1berry/postgres
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
postgres/src/backend/nodes/nodeFuncs.c

3878 lines
105 KiB
C
Raw Normal View History

Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
/*-------------------------------------------------------------------------
*
Change my-function-name-- to my_function_name, and optimizer renames.
1999-02-13 23:22:53 +00:00
* nodeFuncs.c
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
* Various general-purpose manipulations of Node trees
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
*
Update copyright for 2018 Backpatch-through: certain files through 9.3
2018-01-02 23:30:12 -05:00
* Portions Copyright (c) 1996-2018, PostgreSQL Global Development Group
Add: * Portions Copyright (c) 1996-2000, PostgreSQL, Inc to all files copyright Regents of Berkeley. Man, that's a lot of files.
2000-01-26 05:58:53 +00:00
* Portions Copyright (c) 1994, Regents of the University of California
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
*
*
* IDENTIFICATION
Remove cvs keywords from all files.
2010-09-20 22:08:53 +02:00
* src/backend/nodes/nodeFuncs.c
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
*
*-------------------------------------------------------------------------
*/
More Makefile cleanups Add #include "postgres.h" as required
1996-10-31 10:42:56 +00:00
#include "postgres.h"
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
#include "catalog/pg_collation.h"
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
#include "catalog/pg_type.h"
#include "miscadmin.h"
Code review for protransform patches. Fix loss of previous expression-simplification work when a transform function fires: we must not simply revert to untransformed input tree. Instead build a dummy FuncExpr node to pass to the transform function. This has the additional advantage of providing a simpler, more uniform API for transform functions. Move documentation to a somewhat less buried spot, relocate some poorly-placed code, be more wary of null constants and invalid typmod values, add an opr_sanity check on protransform function signatures, and some other minor cosmetic adjustments. Note: although this patch touches pg_proc.h, no need for catversion bump, because the changes are cosmetic and don't actually change the intended catalog contents.
2012-03-23 17:29:57 -04:00
#include "nodes/makefuncs.h"
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
#include "nodes/execnodes.h"
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
#include "nodes/nodeFuncs.h"
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
#include "nodes/relation.h"
#include "utils/builtins.h"
#include "utils/lsyscache.h"
Support expressions of the form 'scalar op ANY (array)' and 'scalar op ALL (array)', where the operator is applied between the lefthand scalar and each element of the array. The operator must yield boolean; the result of the construct is the OR or AND of the per-element results, respectively. Original coding by Joe Conway, after an idea of Peter's. Rewritten by Tom to keep the implementation strictly separate from subqueries.
2003-06-29 00:33:44 +00:00
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
static bool expression_returns_set_walker(Node *node, void *context);
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
static int leftmostLoc(int loc1, int loc2);
Refactor to reduce code duplication for function property checking. As noted by Andres Freund, we'd accumulated quite a few similar functions in clauses.c that examine all functions in an expression tree to see if they satisfy some boolean test. Reduce the duplication by inventing a function check_functions_in_node() that applies a simple callback function to each SQL function OID appearing in a given expression node. This also fixes some arguable oversights; for example, contain_mutable_functions() did not check aggregate or window functions for mutability. I doubt that that represents a live bug at the moment, because we don't really consider mutability for aggregates; but it might someday be one. I chose to put check_functions_in_node() in nodeFuncs.c because it seemed like other modules might wish to use it in future. That in turn forced moving set_opfuncid() et al into nodeFuncs.c, as the alternative was for nodeFuncs.c to depend on optimizer/setrefs.c which didn't seem very clean. In passing, teach contain_leaked_vars_walker() about a few more expression node types it can safely look through, and improve the rather messy and undercommented code in has_parallel_hazard_walker(). Discussion: <20160527185853.ziol2os2zskahl7v@alap3.anarazel.de>
2016-06-10 16:03:37 -04:00
static bool fix_opfuncids_walker(Node *node, void *context);
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
static bool planstate_walk_subplans(List *plans, bool (*walker) (),
Initial pgindent run with pg_bsd_indent version 2.0. The new indent version includes numerous fixes thanks to Piotr Stefaniak. The main changes visible in this commit are: * Nicer formatting of function-pointer declarations. * No longer unexpectedly removes spaces in expressions using casts, sizeof, or offsetof. * No longer wants to add a space in "struct structname *varname", as well as some similar cases for const- or volatile-qualified pointers. * Declarations using PG_USED_FOR_ASSERTS_ONLY are formatted more nicely. * Fixes bug where comments following declarations were sometimes placed with no space separating them from the code. * Fixes some odd decisions for comments following case labels. * Fixes some cases where comments following code were indented to less than the expected column 33. On the less good side, it now tends to put more whitespace around typedef names that are not listed in typedefs.list. This might encourage us to put more effort into typedef name collection; it's not really a bug in indent itself. There are more changes coming after this round, having to do with comment indentation and alignment of lines appearing within parentheses. I wanted to limit the size of the diffs to something that could be reviewed without one's eyes completely glazing over, so it seemed better to split up the changes as much as practical. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 14:39:04 -04:00
void *context);
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
static bool planstate_walk_members(List *plans, PlanState **planstates,
bool (*walker) (), void *context);
Make functions static where possible, enclose unused functions in #ifdef NOT_USED.
1997-08-19 21:40:56 +00:00
Phase 2 of read-only-plans project: restructure expression-tree nodes so that all executable expression nodes inherit from a common supertype Expr. This is somewhat of an exercise in code purity rather than any real functional advance, but getting rid of the extra Oper or Func node formerly used in each operator or function call should provide at least a little space and speed improvement. initdb forced by changes in stored-rules representation.
2002-12-12 15:49:42 +00:00
Massive commit to run PGINDENT on all *.c and *.h files.
1997-09-07 05:04:48 +00:00
/*
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
* exprType -
Make SubPlan nodes carry the result's typmod as well as datatype OID. This is for consistency with the (relatively) recent addition of typmod to SubLink. An example of why it's a good idea is to be seen in the recent "failed to locate grouping columns" bug, which wouldn't have happened if a SubPlan exposed the same typmod info as the SubLink it was derived from. This could be back-patched, since it doesn't affect any on-disk data format, but for the moment it doesn't seem necessary to do so.
2009-03-10 22:09:26 +00:00
* returns the Oid of the type of the expression's result.
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
*/
Oid
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
exprType(const Node *expr)
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
{
Oid type;
if (!expr)
return InvalidOid;
switch (nodeTag(expr))
{
case T_Var:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const Var *) expr)->vartype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_Const:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const Const *) expr)->consttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_Param:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const Param *) expr)->paramtype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_Aggref:
Rethink node-level representation of partial-aggregation modes. The original coding had three separate booleans representing partial aggregation behavior, which was confusing, unreadable, and error-prone, not least because the booleans weren't always listed in the same order. It was also inadequate for the allegedly-desirable future extension to support intermediate partial aggregation, because we'd need separate markers for serialization and deserialization in such a case. Merge these bools into an enum "AggSplit" to provide symbolic names for the supported operating modes (and document what those are). By assigning the values of the enum constants carefully, we can treat AggSplit values as options bitmasks so that tests of what to do aren't noticeably more expensive than before. While at it, get rid of Aggref.aggoutputtype. That's not needed since commit 59a3795c2 got rid of setrefs.c's special-purpose Aggref comparison code, and it likewise seemed more confusing than helpful. Assorted comment cleanup as well (there's still more that I want to do in that line). catversion bump for change in Aggref node contents. Should be the last one for partial-aggregation changes. Discussion: <29309.1466699160@sss.pgh.pa.us>
2016-06-26 14:33:38 -04:00
type = ((const Aggref *) expr)->aggtype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingFunc:
type = INT4OID;
break;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
case T_WindowFunc:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const WindowFunc *) expr)->wintype;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ArrayRef:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const ArrayRef *arrayref = (const ArrayRef *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
/* slice and/or store operations yield the array type */
if (arrayref->reflowerindexpr || arrayref->refassgnexpr)
type = arrayref->refarraytype;
else
type = arrayref->refelemtype;
}
break;
case T_FuncExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const FuncExpr *) expr)->funcresulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
Support use of function argument names to identify which actual arguments match which function parameters. The syntax uses AS, for example funcname(value AS arg1, anothervalue AS arg2) Pavel Stehule
2009-10-08 02:39:25 +00:00
case T_NamedArgExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = exprType((Node *) ((const NamedArgExpr *) expr)->arg);
Support use of function argument names to identify which actual arguments match which function parameters. The syntax uses AS, for example funcname(value AS arg1, anothervalue AS arg2) Pavel Stehule
2009-10-08 02:39:25 +00:00
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_OpExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const OpExpr *) expr)->opresulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_DistinctExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const DistinctExpr *) expr)->opresulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_NullIfExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const NullIfExpr *) expr)->opresulttype;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ScalarArrayOpExpr:
type = BOOLOID;
break;
case T_BoolExpr:
type = BOOLOID;
break;
case T_SubLink:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const SubLink *sublink = (const SubLink *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (sublink->subLinkType == EXPR_SUBLINK ||
sublink->subLinkType == ARRAY_SUBLINK)
{
/* get the type of the subselect's first target column */
Query *qtree = (Query *) sublink->subselect;
TargetEntry *tent;
if (!qtree || !IsA(qtree, Query))
elog(ERROR, "cannot get type for untransformed sublink");
Improve castNode notation by introducing list-extraction-specific variants. This extends the castNode() notation introduced by commit 5bcab1114 to provide, in one step, extraction of a list cell's pointer and coercion to a concrete node type. For example, "lfirst_node(Foo, lc)" is the same as "castNode(Foo, lfirst(lc))". Almost half of the uses of castNode that have appeared so far include a list extraction call, so this is pretty widely useful, and it saves a few more keystrokes compared to the old way. As with the previous patch, back-patch the addition of these macros to pg_list.h, so that the notation will be available when back-patching. Patch by me, after an idea of Andrew Gierth's. Discussion: https://postgr.es/m/14197.1491841216@sss.pgh.pa.us
2017-04-10 13:51:29 -04:00
tent = linitial_node(TargetEntry, qtree->targetList);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Assert(!tent->resjunk);
type = exprType((Node *) tent->expr);
if (sublink->subLinkType == ARRAY_SUBLINK)
{
Support arrays as input to array_agg() and ARRAY(SELECT ...). These cases formerly failed with errors about "could not find array type for data type". Now they yield arrays of the same element type and one higher dimension. The implementation involves creating functions with API similar to the existing accumArrayResult() family. I (tgl) also extended the base family by adding an initArrayResult() function, which allows callers to avoid special-casing the zero-inputs case if they just want an empty array as result. (Not all do, so the previous calling convention remains valid.) This allowed simplifying some existing code in xml.c and plperl.c. Ali Akbar, reviewed by Pavel Stehule, significantly modified by me
2014-11-25 12:21:22 -05:00
type = get_promoted_array_type(type);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (!OidIsValid(type))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("could not find array type for data type %s",
Phase 3 of pgindent updates. Don't move parenthesized lines to the left, even if that means they flow past the right margin. By default, BSD indent lines up statement continuation lines that are within parentheses so that they start just to the right of the preceding left parenthesis. However, traditionally, if that resulted in the continuation line extending to the right of the desired right margin, then indent would push it left just far enough to not overrun the margin, if it could do so without making the continuation line start to the left of the current statement indent. That makes for a weird mix of indentations unless one has been completely rigid about never violating the 80-column limit. This behavior has been pretty universally panned by Postgres developers. Hence, disable it with indent's new -lpl switch, so that parenthesized lines are always lined up with the preceding left paren. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:35:54 -04:00
format_type_be(exprType((Node *) tent->expr)))));
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
}
}
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
else if (sublink->subLinkType == MULTIEXPR_SUBLINK)
{
/* MULTIEXPR is always considered to return RECORD */
type = RECORDOID;
}
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
else
{
/* for all other sublink types, result is boolean */
type = BOOLOID;
}
}
break;
case T_SubPlan:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const SubPlan *subplan = (const SubPlan *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (subplan->subLinkType == EXPR_SUBLINK ||
subplan->subLinkType == ARRAY_SUBLINK)
{
/* get the type of the subselect's first target column */
type = subplan->firstColType;
if (subplan->subLinkType == ARRAY_SUBLINK)
{
Support arrays as input to array_agg() and ARRAY(SELECT ...). These cases formerly failed with errors about "could not find array type for data type". Now they yield arrays of the same element type and one higher dimension. The implementation involves creating functions with API similar to the existing accumArrayResult() family. I (tgl) also extended the base family by adding an initArrayResult() function, which allows callers to avoid special-casing the zero-inputs case if they just want an empty array as result. (Not all do, so the previous calling convention remains valid.) This allowed simplifying some existing code in xml.c and plperl.c. Ali Akbar, reviewed by Pavel Stehule, significantly modified by me
2014-11-25 12:21:22 -05:00
type = get_promoted_array_type(type);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (!OidIsValid(type))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("could not find array type for data type %s",
Phase 3 of pgindent updates. Don't move parenthesized lines to the left, even if that means they flow past the right margin. By default, BSD indent lines up statement continuation lines that are within parentheses so that they start just to the right of the preceding left parenthesis. However, traditionally, if that resulted in the continuation line extending to the right of the desired right margin, then indent would push it left just far enough to not overrun the margin, if it could do so without making the continuation line start to the left of the current statement indent. That makes for a weird mix of indentations unless one has been completely rigid about never violating the 80-column limit. This behavior has been pretty universally panned by Postgres developers. Hence, disable it with indent's new -lpl switch, so that parenthesized lines are always lined up with the preceding left paren. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:35:54 -04:00
format_type_be(subplan->firstColType))));
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
}
}
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
else if (subplan->subLinkType == MULTIEXPR_SUBLINK)
{
/* MULTIEXPR is always considered to return RECORD */
type = RECORDOID;
}
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
else
{
/* for all other subplan types, result is boolean */
type = BOOLOID;
}
}
break;
case T_AlternativeSubPlan:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const AlternativeSubPlan *asplan = (const AlternativeSubPlan *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
/* subplans should all return the same thing */
type = exprType((Node *) linitial(asplan->subplans));
}
break;
case T_FieldSelect:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const FieldSelect *) expr)->resulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_FieldStore:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const FieldStore *) expr)->resulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_RelabelType:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const RelabelType *) expr)->resulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_CoerceViaIO:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const CoerceViaIO *) expr)->resulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_ArrayCoerceExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const ArrayCoerceExpr *) expr)->resulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_ConvertRowtypeExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const ConvertRowtypeExpr *) expr)->resulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
case T_CollateExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = exprType((Node *) ((const CollateExpr *) expr)->arg);
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_CaseExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const CaseExpr *) expr)->casetype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_CaseTestExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const CaseTestExpr *) expr)->typeId;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_ArrayExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const ArrayExpr *) expr)->array_typeid;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_RowExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const RowExpr *) expr)->row_typeid;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_RowCompareExpr:
type = BOOLOID;
break;
case T_CoalesceExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const CoalesceExpr *) expr)->coalescetype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_MinMaxExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const MinMaxExpr *) expr)->minmaxtype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
Improve parsetree representation of special functions such as CURRENT_DATE. We implement a dozen or so parameterless functions that the SQL standard defines special syntax for. Up to now, that was done by converting them into more or less ad-hoc constructs such as "'now'::text::date". That's messy for multiple reasons: it exposes what should be implementation details to users, and performance is worse than it needs to be in several cases. To improve matters, invent a new expression node type SQLValueFunction that can represent any of these parameterless functions. Bump catversion because this changes stored parsetrees for rules. Discussion: <30058.1463091294@sss.pgh.pa.us>
2016-08-16 20:33:01 -04:00
case T_SQLValueFunction:
type = ((const SQLValueFunction *) expr)->type;
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_XmlExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
if (((const XmlExpr *) expr)->op == IS_DOCUMENT)
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
type = BOOLOID;
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
else if (((const XmlExpr *) expr)->op == IS_XMLSERIALIZE)
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
type = TEXTOID;
else
type = XMLOID;
break;
case T_NullTest:
type = BOOLOID;
break;
case T_BooleanTest:
type = BOOLOID;
break;
case T_CoerceToDomain:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const CoerceToDomain *) expr)->resulttype;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_CoerceToDomainValue:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const CoerceToDomainValue *) expr)->typeId;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_SetToDefault:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = ((const SetToDefault *) expr)->typeId;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case T_CurrentOfExpr:
type = BOOLOID;
break;
Identity columns This is the SQL standard-conforming variant of PostgreSQL's serial columns. It fixes a few usability issues that serial columns have: - CREATE TABLE / LIKE copies default but refers to same sequence - cannot add/drop serialness with ALTER TABLE - dropping default does not drop sequence - need to grant separate privileges to sequence - other slight weirdnesses because serial is some kind of special macro Reviewed-by: Vitaly Burovoy <vitaly.burovoy@gmail.com>
2017-04-06 08:33:16 -04:00
case T_NextValueExpr:
type = ((const NextValueExpr *) expr)->typeId;
break;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_InferenceElem:
{
const InferenceElem *n = (const InferenceElem *) expr;
type = exprType((Node *) n->expr);
}
break;
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
case T_PlaceHolderVar:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
type = exprType((Node *) ((const PlaceHolderVar *) expr)->phexpr);
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
default:
elog(ERROR, "unrecognized node type: %d", (int) nodeTag(expr));
type = InvalidOid; /* keep compiler quiet */
break;
}
return type;
}
/*
* exprTypmod -
Make SubPlan nodes carry the result's typmod as well as datatype OID. This is for consistency with the (relatively) recent addition of typmod to SubLink. An example of why it's a good idea is to be seen in the recent "failed to locate grouping columns" bug, which wouldn't have happened if a SubPlan exposed the same typmod info as the SubLink it was derived from. This could be back-patched, since it doesn't affect any on-disk data format, but for the moment it doesn't seem necessary to do so.
2009-03-10 22:09:26 +00:00
* returns the type-specific modifier of the expression's result type,
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 12:12:18 -04:00
* if it can be determined. In many cases, it can't and we return -1.
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
*/
int32
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
exprTypmod(const Node *expr)
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
{
if (!expr)
return -1;
switch (nodeTag(expr))
{
case T_Var:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const Var *) expr)->vartypmod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_Const:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const Const *) expr)->consttypmod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_Param:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const Param *) expr)->paramtypmod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ArrayRef:
/* typmod is the same for array or element */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const ArrayRef *) expr)->reftypmod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_FuncExpr:
{
int32 coercedTypmod;
/* Be smart about length-coercion functions... */
if (exprIsLengthCoercion(expr, &coercedTypmod))
return coercedTypmod;
}
break;
Support use of function argument names to identify which actual arguments match which function parameters. The syntax uses AS, for example funcname(value AS arg1, anothervalue AS arg2) Pavel Stehule
2009-10-08 02:39:25 +00:00
case T_NamedArgExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return exprTypmod((Node *) ((const NamedArgExpr *) expr)->arg);
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_NullIfExpr:
{
/*
* Result is either first argument or NULL, so we can report
* first argument's typmod if known.
*/
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const NullIfExpr *nexpr = (const NullIfExpr *) expr;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
return exprTypmod((Node *) linitial(nexpr->args));
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_SubLink:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const SubLink *sublink = (const SubLink *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (sublink->subLinkType == EXPR_SUBLINK ||
sublink->subLinkType == ARRAY_SUBLINK)
{
/* get the typmod of the subselect's first target column */
Query *qtree = (Query *) sublink->subselect;
TargetEntry *tent;
if (!qtree || !IsA(qtree, Query))
elog(ERROR, "cannot get type for untransformed sublink");
Improve castNode notation by introducing list-extraction-specific variants. This extends the castNode() notation introduced by commit 5bcab1114 to provide, in one step, extraction of a list cell's pointer and coercion to a concrete node type. For example, "lfirst_node(Foo, lc)" is the same as "castNode(Foo, lfirst(lc))". Almost half of the uses of castNode that have appeared so far include a list extraction call, so this is pretty widely useful, and it saves a few more keystrokes compared to the old way. As with the previous patch, back-patch the addition of these macros to pg_list.h, so that the notation will be available when back-patching. Patch by me, after an idea of Andrew Gierth's. Discussion: https://postgr.es/m/14197.1491841216@sss.pgh.pa.us
2017-04-10 13:51:29 -04:00
tent = linitial_node(TargetEntry, qtree->targetList);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Assert(!tent->resjunk);
return exprTypmod((Node *) tent->expr);
/* note we don't need to care if it's an array */
}
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
/* otherwise, result is RECORD or BOOLEAN, typmod is -1 */
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
}
break;
Make SubPlan nodes carry the result's typmod as well as datatype OID. This is for consistency with the (relatively) recent addition of typmod to SubLink. An example of why it's a good idea is to be seen in the recent "failed to locate grouping columns" bug, which wouldn't have happened if a SubPlan exposed the same typmod info as the SubLink it was derived from. This could be back-patched, since it doesn't affect any on-disk data format, but for the moment it doesn't seem necessary to do so.
2009-03-10 22:09:26 +00:00
case T_SubPlan:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const SubPlan *subplan = (const SubPlan *) expr;
Make SubPlan nodes carry the result's typmod as well as datatype OID. This is for consistency with the (relatively) recent addition of typmod to SubLink. An example of why it's a good idea is to be seen in the recent "failed to locate grouping columns" bug, which wouldn't have happened if a SubPlan exposed the same typmod info as the SubLink it was derived from. This could be back-patched, since it doesn't affect any on-disk data format, but for the moment it doesn't seem necessary to do so.
2009-03-10 22:09:26 +00:00
if (subplan->subLinkType == EXPR_SUBLINK ||
subplan->subLinkType == ARRAY_SUBLINK)
{
/* get the typmod of the subselect's first target column */
/* note we don't need to care if it's an array */
return subplan->firstColTypmod;
}
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
/* otherwise, result is RECORD or BOOLEAN, typmod is -1 */
Make SubPlan nodes carry the result's typmod as well as datatype OID. This is for consistency with the (relatively) recent addition of typmod to SubLink. An example of why it's a good idea is to be seen in the recent "failed to locate grouping columns" bug, which wouldn't have happened if a SubPlan exposed the same typmod info as the SubLink it was derived from. This could be back-patched, since it doesn't affect any on-disk data format, but for the moment it doesn't seem necessary to do so.
2009-03-10 22:09:26 +00:00
}
break;
case T_AlternativeSubPlan:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const AlternativeSubPlan *asplan = (const AlternativeSubPlan *) expr;
Make SubPlan nodes carry the result's typmod as well as datatype OID. This is for consistency with the (relatively) recent addition of typmod to SubLink. An example of why it's a good idea is to be seen in the recent "failed to locate grouping columns" bug, which wouldn't have happened if a SubPlan exposed the same typmod info as the SubLink it was derived from. This could be back-patched, since it doesn't affect any on-disk data format, but for the moment it doesn't seem necessary to do so.
2009-03-10 22:09:26 +00:00
/* subplans should all return the same thing */
return exprTypmod((Node *) linitial(asplan->subplans));
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_FieldSelect:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const FieldSelect *) expr)->resulttypmod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_RelabelType:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const RelabelType *) expr)->resulttypmod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ArrayCoerceExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const ArrayCoerceExpr *) expr)->resulttypmod;
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
case T_CollateExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return exprTypmod((Node *) ((const CollateExpr *) expr)->arg);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_CaseExpr:
{
/*
* If all the alternatives agree on type/typmod, return that
* typmod, else use -1
*/
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const CaseExpr *cexpr = (const CaseExpr *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Oid casetype = cexpr->casetype;
int32 typmod;
ListCell *arg;
if (!cexpr->defresult)
return -1;
if (exprType((Node *) cexpr->defresult) != casetype)
return -1;
typmod = exprTypmod((Node *) cexpr->defresult);
if (typmod < 0)
return -1; /* no point in trying harder */
foreach(arg, cexpr->args)
{
Improve castNode notation by introducing list-extraction-specific variants. This extends the castNode() notation introduced by commit 5bcab1114 to provide, in one step, extraction of a list cell's pointer and coercion to a concrete node type. For example, "lfirst_node(Foo, lc)" is the same as "castNode(Foo, lfirst(lc))". Almost half of the uses of castNode that have appeared so far include a list extraction call, so this is pretty widely useful, and it saves a few more keystrokes compared to the old way. As with the previous patch, back-patch the addition of these macros to pg_list.h, so that the notation will be available when back-patching. Patch by me, after an idea of Andrew Gierth's. Discussion: https://postgr.es/m/14197.1491841216@sss.pgh.pa.us
2017-04-10 13:51:29 -04:00
CaseWhen *w = lfirst_node(CaseWhen, arg);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (exprType((Node *) w->result) != casetype)
return -1;
if (exprTypmod((Node *) w->result) != typmod)
return -1;
}
return typmod;
}
break;
case T_CaseTestExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const CaseTestExpr *) expr)->typeMod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ArrayExpr:
{
/*
* If all the elements agree on type/typmod, return that
* typmod, else use -1
*/
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const ArrayExpr *arrayexpr = (const ArrayExpr *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Oid commontype;
int32 typmod;
ListCell *elem;
if (arrayexpr->elements == NIL)
return -1;
typmod = exprTypmod((Node *) linitial(arrayexpr->elements));
if (typmod < 0)
return -1; /* no point in trying harder */
if (arrayexpr->multidims)
commontype = arrayexpr->array_typeid;
else
commontype = arrayexpr->element_typeid;
foreach(elem, arrayexpr->elements)
{
Node *e = (Node *) lfirst(elem);
if (exprType(e) != commontype)
return -1;
if (exprTypmod(e) != typmod)
return -1;
}
return typmod;
}
break;
case T_CoalesceExpr:
{
/*
* If all the alternatives agree on type/typmod, return that
* typmod, else use -1
*/
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const CoalesceExpr *cexpr = (const CoalesceExpr *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Oid coalescetype = cexpr->coalescetype;
int32 typmod;
ListCell *arg;
if (exprType((Node *) linitial(cexpr->args)) != coalescetype)
return -1;
typmod = exprTypmod((Node *) linitial(cexpr->args));
if (typmod < 0)
return -1; /* no point in trying harder */
for_each_cell(arg, lnext(list_head(cexpr->args)))
{
Node *e = (Node *) lfirst(arg);
if (exprType(e) != coalescetype)
return -1;
if (exprTypmod(e) != typmod)
return -1;
}
return typmod;
}
break;
case T_MinMaxExpr:
{
/*
* If all the alternatives agree on type/typmod, return that
* typmod, else use -1
*/
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const MinMaxExpr *mexpr = (const MinMaxExpr *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Oid minmaxtype = mexpr->minmaxtype;
int32 typmod;
ListCell *arg;
if (exprType((Node *) linitial(mexpr->args)) != minmaxtype)
return -1;
typmod = exprTypmod((Node *) linitial(mexpr->args));
if (typmod < 0)
return -1; /* no point in trying harder */
for_each_cell(arg, lnext(list_head(mexpr->args)))
{
Node *e = (Node *) lfirst(arg);
if (exprType(e) != minmaxtype)
return -1;
if (exprTypmod(e) != typmod)
return -1;
}
return typmod;
}
break;
Improve parsetree representation of special functions such as CURRENT_DATE. We implement a dozen or so parameterless functions that the SQL standard defines special syntax for. Up to now, that was done by converting them into more or less ad-hoc constructs such as "'now'::text::date". That's messy for multiple reasons: it exposes what should be implementation details to users, and performance is worse than it needs to be in several cases. To improve matters, invent a new expression node type SQLValueFunction that can represent any of these parameterless functions. Bump catversion because this changes stored parsetrees for rules. Discussion: <30058.1463091294@sss.pgh.pa.us>
2016-08-16 20:33:01 -04:00
case T_SQLValueFunction:
return ((const SQLValueFunction *) expr)->typmod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_CoerceToDomain:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const CoerceToDomain *) expr)->resulttypmod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_CoerceToDomainValue:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const CoerceToDomainValue *) expr)->typeMod;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_SetToDefault:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return ((const SetToDefault *) expr)->typeMod;
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
case T_PlaceHolderVar:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
return exprTypmod((Node *) ((const PlaceHolderVar *) expr)->phexpr);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
default:
break;
}
return -1;
}
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
/*
* exprIsLengthCoercion
* Detect whether an expression tree is an application of a datatype's
* typmod-coercion function. Optionally extract the result's typmod.
*
* If coercedTypmod is not NULL, the typmod is stored there if the expression
* is a length-coercion function, else -1 is stored there.
*
* Note that a combined type-and-length coercion will be treated as a
* length coercion by this routine.
*/
bool
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
exprIsLengthCoercion(const Node *expr, int32 *coercedTypmod)
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
{
if (coercedTypmod != NULL)
*coercedTypmod = -1; /* default result on failure */
/*
* Scalar-type length coercions are FuncExprs, array-type length coercions
* are ArrayCoerceExprs
*/
if (expr && IsA(expr, FuncExpr))
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const FuncExpr *func = (const FuncExpr *) expr;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
int nargs;
Const *second_arg;
/*
* If it didn't come from a coercion context, reject.
*/
if (func->funcformat != COERCE_EXPLICIT_CAST &&
func->funcformat != COERCE_IMPLICIT_CAST)
return false;
/*
* If it's not a two-argument or three-argument function with the
* second argument being an int4 constant, it can't have been created
* from a length coercion (it must be a type coercion, instead).
*/
nargs = list_length(func->args);
if (nargs < 2 || nargs > 3)
return false;
second_arg = (Const *) lsecond(func->args);
if (!IsA(second_arg, Const) ||
second_arg->consttype != INT4OID ||
second_arg->constisnull)
return false;
/*
* OK, it is indeed a length-coercion function.
*/
if (coercedTypmod != NULL)
*coercedTypmod = DatumGetInt32(second_arg->constvalue);
return true;
}
if (expr && IsA(expr, ArrayCoerceExpr))
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const ArrayCoerceExpr *acoerce = (const ArrayCoerceExpr *) expr;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
/* It's not a length coercion unless there's a nondefault typmod */
if (acoerce->resulttypmod < 0)
return false;
/*
* OK, it is indeed a length-coercion expression.
*/
if (coercedTypmod != NULL)
*coercedTypmod = acoerce->resulttypmod;
return true;
}
return false;
}
Code review for protransform patches. Fix loss of previous expression-simplification work when a transform function fires: we must not simply revert to untransformed input tree. Instead build a dummy FuncExpr node to pass to the transform function. This has the additional advantage of providing a simpler, more uniform API for transform functions. Move documentation to a somewhat less buried spot, relocate some poorly-placed code, be more wary of null constants and invalid typmod values, add an opr_sanity check on protransform function signatures, and some other minor cosmetic adjustments. Note: although this patch touches pg_proc.h, no need for catversion bump, because the changes are cosmetic and don't actually change the intended catalog contents.
2012-03-23 17:29:57 -04:00
/*
* relabel_to_typmod
* Add a RelabelType node that changes just the typmod of the expression.
*
* This is primarily intended to be used during planning. Therefore, it
* strips any existing RelabelType nodes to maintain the planner's invariant
Get rid of COERCE_DONTCARE. We don't need this hack any more.
2012-10-12 13:35:00 -04:00
* that there are not adjacent RelabelTypes.
Code review for protransform patches. Fix loss of previous expression-simplification work when a transform function fires: we must not simply revert to untransformed input tree. Instead build a dummy FuncExpr node to pass to the transform function. This has the additional advantage of providing a simpler, more uniform API for transform functions. Move documentation to a somewhat less buried spot, relocate some poorly-placed code, be more wary of null constants and invalid typmod values, add an opr_sanity check on protransform function signatures, and some other minor cosmetic adjustments. Note: although this patch touches pg_proc.h, no need for catversion bump, because the changes are cosmetic and don't actually change the intended catalog contents.
2012-03-23 17:29:57 -04:00
*/
Node *
relabel_to_typmod(Node *expr, int32 typmod)
{
Oid type = exprType(expr);
Oid coll = exprCollation(expr);
/* Strip any existing RelabelType node(s) */
while (expr && IsA(expr, RelabelType))
expr = (Node *) ((RelabelType *) expr)->arg;
/* Apply new typmod, preserving the previous exposed type and collation */
return (Node *) makeRelabelType((Expr *) expr, type, typmod, coll,
Get rid of COERCE_DONTCARE. We don't need this hack any more.
2012-10-12 13:35:00 -04:00
COERCE_EXPLICIT_CAST);
Code review for protransform patches. Fix loss of previous expression-simplification work when a transform function fires: we must not simply revert to untransformed input tree. Instead build a dummy FuncExpr node to pass to the transform function. This has the additional advantage of providing a simpler, more uniform API for transform functions. Move documentation to a somewhat less buried spot, relocate some poorly-placed code, be more wary of null constants and invalid typmod values, add an opr_sanity check on protransform function signatures, and some other minor cosmetic adjustments. Note: although this patch touches pg_proc.h, no need for catversion bump, because the changes are cosmetic and don't actually change the intended catalog contents.
2012-03-23 17:29:57 -04:00
}
Move strip_implicit_coercions() from optimizer to nodeFuncs.c. Use of this function has spread into the parser and rewriter, so it seems like time to pull it out of the optimizer and put it into the more central nodeFuncs module. This eliminates the need to #include optimizer/clauses.h in most of the calling files, demonstrating that this function was indeed a bit outside the normal code reference patterns.
2013-07-23 18:21:19 -04:00
/*
* strip_implicit_coercions: remove implicit coercions at top level of tree
*
* This doesn't modify or copy the input expression tree, just return a
* pointer to a suitable place within it.
*
* Note: there isn't any useful thing we can do with a RowExpr here, so
* just return it unchanged, even if it's marked as an implicit coercion.
*/
Node *
strip_implicit_coercions(Node *node)
{
if (node == NULL)
return NULL;
if (IsA(node, FuncExpr))
{
FuncExpr *f = (FuncExpr *) node;
if (f->funcformat == COERCE_IMPLICIT_CAST)
return strip_implicit_coercions(linitial(f->args));
}
else if (IsA(node, RelabelType))
{
RelabelType *r = (RelabelType *) node;
if (r->relabelformat == COERCE_IMPLICIT_CAST)
return strip_implicit_coercions((Node *) r->arg);
}
else if (IsA(node, CoerceViaIO))
{
CoerceViaIO *c = (CoerceViaIO *) node;
if (c->coerceformat == COERCE_IMPLICIT_CAST)
return strip_implicit_coercions((Node *) c->arg);
}
else if (IsA(node, ArrayCoerceExpr))
{
ArrayCoerceExpr *c = (ArrayCoerceExpr *) node;
if (c->coerceformat == COERCE_IMPLICIT_CAST)
return strip_implicit_coercions((Node *) c->arg);
}
else if (IsA(node, ConvertRowtypeExpr))
{
ConvertRowtypeExpr *c = (ConvertRowtypeExpr *) node;
if (c->convertformat == COERCE_IMPLICIT_CAST)
return strip_implicit_coercions((Node *) c->arg);
}
else if (IsA(node, CoerceToDomain))
{
CoerceToDomain *c = (CoerceToDomain *) node;
if (c->coercionformat == COERCE_IMPLICIT_CAST)
return strip_implicit_coercions((Node *) c->arg);
}
return node;
}
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
/*
* expression_returns_set
* Test whether an expression returns a set result.
*
* Because we use expression_tree_walker(), this can also be applied to
Change TRUE/FALSE to true/false The lower case spellings are C and C++ standard and are used in most parts of the PostgreSQL sources. The upper case spellings are only used in some files/modules. So standardize on the standard spellings. The APIs for ICU, Perl, and Windows define their own TRUE and FALSE, so those are left as is when using those APIs. In code comments, we use the lower-case spelling for the C concepts and keep the upper-case spelling for the SQL concepts. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
2017-08-16 00:22:32 -04:00
* whole targetlists; it'll produce true if any one of the tlist items
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
* returns a set.
*/
bool
expression_returns_set(Node *clause)
{
return expression_returns_set_walker(clause, NULL);
}
static bool
expression_returns_set_walker(Node *node, void *context)
{
if (node == NULL)
return false;
if (IsA(node, FuncExpr))
{
FuncExpr *expr = (FuncExpr *) node;
if (expr->funcretset)
return true;
/* else fall through to check args */
}
if (IsA(node, OpExpr))
{
OpExpr *expr = (OpExpr *) node;
if (expr->opretset)
return true;
/* else fall through to check args */
}
Fix no-longer-valid shortcuts in expression_returns_set(). expression_returns_set() used to short-circuit its recursion upon seeing certain node types, such as DistinctExpr, that it knew the executor did not support set-valued arguments for. That was never inherent, though, just a reflection of laziness in execQual.c. With the new implementation of SRFs there is no reason to think that any scalar-valued expression node could not have a set-valued subexpression, except for AggRefs and WindowFuncs where we know there is a parser check rejecting it. And indeed, the shortcut causes unexpected failures for cases such as a SRF underneath DistinctExpr, because the planner stops looking for SRFs too soon. Discussion: https://postgr.es/m/5259.1497044025@sss.pgh.pa.us
2017-06-14 11:10:05 -04:00
/* Avoid recursion for some cases that parser checks not to return a set */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
if (IsA(node, Aggref))
return false;
if (IsA(node, WindowFunc))
return false;
return expression_tree_walker(node, expression_returns_set_walker,
context);
}
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
/*
* exprCollation -
* returns the Oid of the collation of the expression's result.
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
*
* Note: expression nodes that can invoke functions generally have an
* "inputcollid" field, which is what the function should use as collation.
* That is the resolved common collation of the node's inputs. It is often
* but not always the same as the result collation; in particular, if the
* function produces a non-collatable result type from collatable inputs
* or vice versa, the two are different.
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
*/
Oid
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
exprCollation(const Node *expr)
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
{
Oid coll;
if (!expr)
return InvalidOid;
switch (nodeTag(expr))
{
case T_Var:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const Var *) expr)->varcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_Const:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const Const *) expr)->constcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_Param:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const Param *) expr)->paramcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_Aggref:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const Aggref *) expr)->aggcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingFunc:
coll = InvalidOid;
break;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
case T_WindowFunc:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const WindowFunc *) expr)->wincollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_ArrayRef:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const ArrayRef *) expr)->refcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_FuncExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const FuncExpr *) expr)->funccollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_NamedArgExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = exprCollation((Node *) ((const NamedArgExpr *) expr)->arg);
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_OpExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const OpExpr *) expr)->opcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_DistinctExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const DistinctExpr *) expr)->opcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_NullIfExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const NullIfExpr *) expr)->opcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_ScalarArrayOpExpr:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always boolean */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_BoolExpr:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always boolean */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_SubLink:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const SubLink *sublink = (const SubLink *) expr;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
if (sublink->subLinkType == EXPR_SUBLINK ||
sublink->subLinkType == ARRAY_SUBLINK)
{
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
/* get the collation of subselect's first target column */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
Query *qtree = (Query *) sublink->subselect;
TargetEntry *tent;
if (!qtree || !IsA(qtree, Query))
elog(ERROR, "cannot get collation for untransformed sublink");
Improve castNode notation by introducing list-extraction-specific variants. This extends the castNode() notation introduced by commit 5bcab1114 to provide, in one step, extraction of a list cell's pointer and coercion to a concrete node type. For example, "lfirst_node(Foo, lc)" is the same as "castNode(Foo, lfirst(lc))". Almost half of the uses of castNode that have appeared so far include a list extraction call, so this is pretty widely useful, and it saves a few more keystrokes compared to the old way. As with the previous patch, back-patch the addition of these macros to pg_list.h, so that the notation will be available when back-patching. Patch by me, after an idea of Andrew Gierth's. Discussion: https://postgr.es/m/14197.1491841216@sss.pgh.pa.us
2017-04-10 13:51:29 -04:00
tent = linitial_node(TargetEntry, qtree->targetList);
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
Assert(!tent->resjunk);
coll = exprCollation((Node *) tent->expr);
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
/* collation doesn't change if it's converted to array */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
}
else
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
{
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
/* otherwise, result is RECORD or BOOLEAN */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
coll = InvalidOid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
}
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
}
break;
case T_SubPlan:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const SubPlan *subplan = (const SubPlan *) expr;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
if (subplan->subLinkType == EXPR_SUBLINK ||
subplan->subLinkType == ARRAY_SUBLINK)
{
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
/* get the collation of subselect's first target column */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
coll = subplan->firstColCollation;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
/* collation doesn't change if it's converted to array */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
}
else
{
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
/* otherwise, result is RECORD or BOOLEAN */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
coll = InvalidOid;
}
}
break;
case T_AlternativeSubPlan:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const AlternativeSubPlan *asplan = (const AlternativeSubPlan *) expr;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
/* subplans should all return the same thing */
coll = exprCollation((Node *) linitial(asplan->subplans));
}
break;
case T_FieldSelect:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const FieldSelect *) expr)->resultcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_FieldStore:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always composite */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_RelabelType:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const RelabelType *) expr)->resultcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_CoerceViaIO:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const CoerceViaIO *) expr)->resultcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_ArrayCoerceExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const ArrayCoerceExpr *) expr)->resultcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_ConvertRowtypeExpr:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always composite */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
case T_CollateExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const CollateExpr *) expr)->collOid;
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
break;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
case T_CaseExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const CaseExpr *) expr)->casecollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_CaseTestExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const CaseTestExpr *) expr)->collation;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_ArrayExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const ArrayExpr *) expr)->array_collid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_RowExpr:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always composite */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_RowCompareExpr:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always boolean */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_CoalesceExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const CoalesceExpr *) expr)->coalescecollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_MinMaxExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const MinMaxExpr *) expr)->minmaxcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
Improve parsetree representation of special functions such as CURRENT_DATE. We implement a dozen or so parameterless functions that the SQL standard defines special syntax for. Up to now, that was done by converting them into more or less ad-hoc constructs such as "'now'::text::date". That's messy for multiple reasons: it exposes what should be implementation details to users, and performance is worse than it needs to be in several cases. To improve matters, invent a new expression node type SQLValueFunction that can represent any of these parameterless functions. Bump catversion because this changes stored parsetrees for rules. Discussion: <30058.1463091294@sss.pgh.pa.us>
2016-08-16 20:33:01 -04:00
case T_SQLValueFunction:
coll = InvalidOid; /* all cases return non-collatable types */
break;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
case T_XmlExpr:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
/*
* XMLSERIALIZE returns text from non-collatable inputs, so its
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
* collation is always default. The other cases return boolean or
* XML, which are non-collatable.
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
*/
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
if (((const XmlExpr *) expr)->op == IS_XMLSERIALIZE)
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
coll = DEFAULT_COLLATION_OID;
else
coll = InvalidOid;
break;
case T_NullTest:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always boolean */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_BooleanTest:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always boolean */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_CoerceToDomain:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const CoerceToDomain *) expr)->resultcollid;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_CoerceToDomainValue:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const CoerceToDomainValue *) expr)->collation;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_SetToDefault:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const SetToDefault *) expr)->collation;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
case T_CurrentOfExpr:
pgindent run before PG 9.1 beta 1.
2011-04-10 11:42:00 -04:00
coll = InvalidOid; /* result is always boolean */
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
Identity columns This is the SQL standard-conforming variant of PostgreSQL's serial columns. It fixes a few usability issues that serial columns have: - CREATE TABLE / LIKE copies default but refers to same sequence - cannot add/drop serialness with ALTER TABLE - dropping default does not drop sequence - need to grant separate privileges to sequence - other slight weirdnesses because serial is some kind of special macro Reviewed-by: Vitaly Burovoy <vitaly.burovoy@gmail.com>
2017-04-06 08:33:16 -04:00
case T_NextValueExpr:
coll = InvalidOid; /* result is always an integer type */
break;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_InferenceElem:
coll = exprCollation((Node *) ((const InferenceElem *) expr)->expr);
break;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
case T_PlaceHolderVar:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = exprCollation((Node *) ((const PlaceHolderVar *) expr)->phexpr);
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
default:
elog(ERROR, "unrecognized node type: %d", (int) nodeTag(expr));
coll = InvalidOid; /* keep compiler quiet */
break;
}
return coll;
}
/*
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
* exprInputCollation -
* returns the Oid of the collation a function should use, if available.
*
* Result is InvalidOid if the node type doesn't store this information.
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
*/
Oid
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
exprInputCollation(const Node *expr)
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
{
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
Oid coll;
if (!expr)
return InvalidOid;
switch (nodeTag(expr))
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
{
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_Aggref:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const Aggref *) expr)->inputcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_WindowFunc:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const WindowFunc *) expr)->inputcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_FuncExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const FuncExpr *) expr)->inputcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_OpExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const OpExpr *) expr)->inputcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_DistinctExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const DistinctExpr *) expr)->inputcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_NullIfExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const NullIfExpr *) expr)->inputcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_ScalarArrayOpExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const ScalarArrayOpExpr *) expr)->inputcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_MinMaxExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
coll = ((const MinMaxExpr *) expr)->inputcollid;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
default:
coll = InvalidOid;
break;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
}
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
return coll;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
}
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
/*
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
* exprSetCollation -
* Assign collation information to an expression tree node.
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
*
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
* Note: since this is only used during parse analysis, we don't need to
* worry about subplans or PlaceHolderVars.
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
*/
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
void
exprSetCollation(Node *expr, Oid collation)
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
{
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
switch (nodeTag(expr))
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
{
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_Var:
((Var *) expr)->varcollid = collation;
break;
case T_Const:
((Const *) expr)->constcollid = collation;
break;
case T_Param:
((Param *) expr)->paramcollid = collation;
break;
case T_Aggref:
((Aggref *) expr)->aggcollid = collation;
break;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingFunc:
Assert(!OidIsValid(collation));
break;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_WindowFunc:
((WindowFunc *) expr)->wincollid = collation;
break;
case T_ArrayRef:
((ArrayRef *) expr)->refcollid = collation;
break;
case T_FuncExpr:
((FuncExpr *) expr)->funccollid = collation;
break;
case T_NamedArgExpr:
Assert(collation == exprCollation((Node *) ((NamedArgExpr *) expr)->arg));
break;
case T_OpExpr:
((OpExpr *) expr)->opcollid = collation;
break;
case T_DistinctExpr:
((DistinctExpr *) expr)->opcollid = collation;
break;
case T_NullIfExpr:
((NullIfExpr *) expr)->opcollid = collation;
break;
case T_ScalarArrayOpExpr:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always boolean */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_BoolExpr:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always boolean */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_SubLink:
#ifdef USE_ASSERT_CHECKING
{
SubLink *sublink = (SubLink *) expr;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
if (sublink->subLinkType == EXPR_SUBLINK ||
sublink->subLinkType == ARRAY_SUBLINK)
{
/* get the collation of subselect's first target column */
Query *qtree = (Query *) sublink->subselect;
TargetEntry *tent;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
if (!qtree || !IsA(qtree, Query))
elog(ERROR, "cannot set collation for untransformed sublink");
Improve castNode notation by introducing list-extraction-specific variants. This extends the castNode() notation introduced by commit 5bcab1114 to provide, in one step, extraction of a list cell's pointer and coercion to a concrete node type. For example, "lfirst_node(Foo, lc)" is the same as "castNode(Foo, lfirst(lc))". Almost half of the uses of castNode that have appeared so far include a list extraction call, so this is pretty widely useful, and it saves a few more keystrokes compared to the old way. As with the previous patch, back-patch the addition of these macros to pg_list.h, so that the notation will be available when back-patching. Patch by me, after an idea of Andrew Gierth's. Discussion: https://postgr.es/m/14197.1491841216@sss.pgh.pa.us
2017-04-10 13:51:29 -04:00
tent = linitial_node(TargetEntry, qtree->targetList);
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
Assert(!tent->resjunk);
Assert(collation == exprCollation((Node *) tent->expr));
}
else
{
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
/* otherwise, result is RECORD or BOOLEAN */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
Assert(!OidIsValid(collation));
}
}
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
#endif /* USE_ASSERT_CHECKING */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_FieldSelect:
((FieldSelect *) expr)->resultcollid = collation;
break;
case T_FieldStore:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always composite */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_RelabelType:
((RelabelType *) expr)->resultcollid = collation;
break;
case T_CoerceViaIO:
((CoerceViaIO *) expr)->resultcollid = collation;
break;
case T_ArrayCoerceExpr:
((ArrayCoerceExpr *) expr)->resultcollid = collation;
break;
case T_ConvertRowtypeExpr:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always composite */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_CaseExpr:
((CaseExpr *) expr)->casecollid = collation;
break;
case T_ArrayExpr:
((ArrayExpr *) expr)->array_collid = collation;
break;
case T_RowExpr:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always composite */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_RowCompareExpr:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always boolean */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_CoalesceExpr:
((CoalesceExpr *) expr)->coalescecollid = collation;
break;
case T_MinMaxExpr:
((MinMaxExpr *) expr)->minmaxcollid = collation;
break;
Improve parsetree representation of special functions such as CURRENT_DATE. We implement a dozen or so parameterless functions that the SQL standard defines special syntax for. Up to now, that was done by converting them into more or less ad-hoc constructs such as "'now'::text::date". That's messy for multiple reasons: it exposes what should be implementation details to users, and performance is worse than it needs to be in several cases. To improve matters, invent a new expression node type SQLValueFunction that can represent any of these parameterless functions. Bump catversion because this changes stored parsetrees for rules. Discussion: <30058.1463091294@sss.pgh.pa.us>
2016-08-16 20:33:01 -04:00
case T_SQLValueFunction:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* no collatable results */
Improve parsetree representation of special functions such as CURRENT_DATE. We implement a dozen or so parameterless functions that the SQL standard defines special syntax for. Up to now, that was done by converting them into more or less ad-hoc constructs such as "'now'::text::date". That's messy for multiple reasons: it exposes what should be implementation details to users, and performance is worse than it needs to be in several cases. To improve matters, invent a new expression node type SQLValueFunction that can represent any of these parameterless functions. Bump catversion because this changes stored parsetrees for rules. Discussion: <30058.1463091294@sss.pgh.pa.us>
2016-08-16 20:33:01 -04:00
break;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_XmlExpr:
Assert((((XmlExpr *) expr)->op == IS_XMLSERIALIZE) ?
(collation == DEFAULT_COLLATION_OID) :
(collation == InvalidOid));
break;
case T_NullTest:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always boolean */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_BooleanTest:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always boolean */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
case T_CoerceToDomain:
((CoerceToDomain *) expr)->resultcollid = collation;
break;
case T_CoerceToDomainValue:
((CoerceToDomainValue *) expr)->collation = collation;
break;
case T_SetToDefault:
((SetToDefault *) expr)->collation = collation;
break;
case T_CurrentOfExpr:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always boolean */
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
break;
Identity columns This is the SQL standard-conforming variant of PostgreSQL's serial columns. It fixes a few usability issues that serial columns have: - CREATE TABLE / LIKE copies default but refers to same sequence - cannot add/drop serialness with ALTER TABLE - dropping default does not drop sequence - need to grant separate privileges to sequence - other slight weirdnesses because serial is some kind of special macro Reviewed-by: Vitaly Burovoy <vitaly.burovoy@gmail.com>
2017-04-06 08:33:16 -04:00
case T_NextValueExpr:
Phase 2 of pgindent updates. Change pg_bsd_indent to follow upstream rules for placement of comments to the right of code, and remove pgindent hack that caused comments following #endif to not obey the general rule. Commit e3860ffa4dd0dad0dd9eea4be9cc1412373a8c89 wasn't actually using the published version of pg_bsd_indent, but a hacked-up version that tried to minimize the amount of movement of comments to the right of code. The situation of interest is where such a comment has to be moved to the right of its default placement at column 33 because there's code there. BSD indent has always moved right in units of tab stops in such cases --- but in the previous incarnation, indent was working in 8-space tab stops, while now it knows we use 4-space tabs. So the net result is that in about half the cases, such comments are placed one tab stop left of before. This is better all around: it leaves more room on the line for comment text, and it means that in such cases the comment uniformly starts at the next 4-space tab stop after the code, rather than sometimes one and sometimes two tabs after. Also, ensure that comments following #endif are indented the same as comments following other preprocessor commands such as #else. That inconsistency turns out to have been self-inflicted damage from a poorly-thought-through post-indent "fixup" in pgindent. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:18:54 -04:00
Assert(!OidIsValid(collation)); /* result is always an integer
* type */
Identity columns This is the SQL standard-conforming variant of PostgreSQL's serial columns. It fixes a few usability issues that serial columns have: - CREATE TABLE / LIKE copies default but refers to same sequence - cannot add/drop serialness with ALTER TABLE - dropping default does not drop sequence - need to grant separate privileges to sequence - other slight weirdnesses because serial is some kind of special macro Reviewed-by: Vitaly Burovoy <vitaly.burovoy@gmail.com>
2017-04-06 08:33:16 -04:00
break;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
default:
elog(ERROR, "unrecognized node type: %d", (int) nodeTag(expr));
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
}
}
/*
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
* exprSetInputCollation -
* Assign input-collation information to an expression tree node.
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
*
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
* This is a no-op for node types that don't store their input collation.
* Note we omit RowCompareExpr, which needs special treatment since it
* contains multiple input collation OIDs.
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
*/
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
void
exprSetInputCollation(Node *expr, Oid inputcollation)
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
{
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
switch (nodeTag(expr))
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
{
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_Aggref:
((Aggref *) expr)->inputcollid = inputcollation;
break;
case T_WindowFunc:
((WindowFunc *) expr)->inputcollid = inputcollation;
break;
case T_FuncExpr:
((FuncExpr *) expr)->inputcollid = inputcollation;
break;
case T_OpExpr:
((OpExpr *) expr)->inputcollid = inputcollation;
break;
case T_DistinctExpr:
((DistinctExpr *) expr)->inputcollid = inputcollation;
break;
case T_NullIfExpr:
((NullIfExpr *) expr)->inputcollid = inputcollation;
break;
case T_ScalarArrayOpExpr:
((ScalarArrayOpExpr *) expr)->inputcollid = inputcollation;
break;
case T_MinMaxExpr:
((MinMaxExpr *) expr)->inputcollid = inputcollation;
break;
default:
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
}
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
}
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/*
* exprLocation -
* returns the parse location of an expression tree, for error reports
*
* -1 is returned if the location can't be determined.
*
* For expressions larger than a single token, the intent here is to
* return the location of the expression's leftmost token, not necessarily
* the topmost Node's location field. For example, an OpExpr's location
* field will point at the operator name, but if it is not a prefix operator
* then we should return the location of the left-hand operand instead.
* The reason is that we want to reference the entire expression not just
* that operator, and pointing to its start seems to be the most natural way.
*
* The location is not perfect --- for example, since the grammar doesn't
* explicitly represent parentheses in the parsetree, given something that
* had been written "(a + b) * c" we are going to point at "a" not "(".
* But it should be plenty good enough for error reporting purposes.
*
* You might think that this code is overly general, for instance why check
* the operands of a FuncExpr node, when the function name can be expected
* to be to the left of them? There are a couple of reasons. The grammar
* sometimes builds expressions that aren't quite what the user wrote;
* for instance x IS NOT BETWEEN ... becomes a NOT-expression whose keyword
* pointer is to the right of its leftmost argument. Also, nodes that were
* inserted implicitly by parse analysis (such as FuncExprs for implicit
* coercions) will have location -1, and so we can have odd combinations of
* known and unknown locations in a tree.
*/
int
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
exprLocation(const Node *expr)
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
{
int loc;
if (expr == NULL)
return -1;
switch (nodeTag(expr))
{
Add a bunch of new error location reports to parse-analysis error messages. There are still some weak spots around JOIN USING and relation alias lists, but most errors reported within backend/parser/ now have locations.
2008-09-01 20:42:46 +00:00
case T_RangeVar:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const RangeVar *) expr)->location;
Add a bunch of new error location reports to parse-analysis error messages. There are still some weak spots around JOIN USING and relation alias lists, but most errors reported within backend/parser/ now have locations.
2008-09-01 20:42:46 +00:00
break;
Support XMLTABLE query expression XMLTABLE is defined by the SQL/XML standard as a feature that allows turning XML-formatted data into relational form, so that it can be used as a <table primary> in the FROM clause of a query. This new construct provides significant simplicity and performance benefit for XML data processing; what in a client-side custom implementation was reported to take 20 minutes can be executed in 400ms using XMLTABLE. (The same functionality was said to take 10 seconds using nested PostgreSQL XPath function calls, and 5 seconds using XMLReader under PL/Python). The implemented syntax deviates slightly from what the standard requires. First, the standard indicates that the PASSING clause is optional and that multiple XML input documents may be given to it; we make it mandatory and accept a single document only. Second, we don't currently support a default namespace to be specified. This implementation relies on a new executor node based on a hardcoded method table. (Because the grammar is fixed, there is no extensibility in the current approach; further constructs can be implemented on top of this such as JSON_TABLE, but they require changes to core code.) Author: Pavel Stehule, Álvaro Herrera Extensively reviewed by: Craig Ringer Discussion: https://postgr.es/m/CAFj8pRAgfzMD-LoSmnMGybD0WsEznLHWap8DO79+-GTRAPR4qA@mail.gmail.com
2017-03-08 12:39:37 -03:00
case T_TableFunc:
loc = ((const TableFunc *) expr)->location;
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_Var:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const Var *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_Const:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const Const *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_Param:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const Param *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_Aggref:
/* function name should always be the first thing */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const Aggref *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingFunc:
loc = ((const GroupingFunc *) expr)->location;
break;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
case T_WindowFunc:
/* function name should always be the first thing */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const WindowFunc *) expr)->location;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_ArrayRef:
/* just use array argument's location */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation((Node *) ((const ArrayRef *) expr)->refexpr);
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_FuncExpr:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const FuncExpr *fexpr = (const FuncExpr *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* consider both function name and leftmost arg */
loc = leftmostLoc(fexpr->location,
exprLocation((Node *) fexpr->args));
}
break;
Support use of function argument names to identify which actual arguments match which function parameters. The syntax uses AS, for example funcname(value AS arg1, anothervalue AS arg2) Pavel Stehule
2009-10-08 02:39:25 +00:00
case T_NamedArgExpr:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const NamedArgExpr *na = (const NamedArgExpr *) expr;
Support use of function argument names to identify which actual arguments match which function parameters. The syntax uses AS, for example funcname(value AS arg1, anothervalue AS arg2) Pavel Stehule
2009-10-08 02:39:25 +00:00
/* consider both argument name and value */
loc = leftmostLoc(na->location,
exprLocation((Node *) na->arg));
}
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_OpExpr:
case T_DistinctExpr: /* struct-equivalent to OpExpr */
case T_NullIfExpr: /* struct-equivalent to OpExpr */
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const OpExpr *opexpr = (const OpExpr *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* consider both operator name and leftmost arg */
loc = leftmostLoc(opexpr->location,
exprLocation((Node *) opexpr->args));
}
break;
case T_ScalarArrayOpExpr:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const ScalarArrayOpExpr *saopexpr = (const ScalarArrayOpExpr *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* consider both operator name and leftmost arg */
loc = leftmostLoc(saopexpr->location,
exprLocation((Node *) saopexpr->args));
}
break;
case T_BoolExpr:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const BoolExpr *bexpr = (const BoolExpr *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/*
* Same as above, to handle either NOT or AND/OR. We can't
* special-case NOT because of the way that it's used for
* things like IS NOT BETWEEN.
*/
loc = leftmostLoc(bexpr->location,
exprLocation((Node *) bexpr->args));
}
break;
case T_SubLink:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const SubLink *sublink = (const SubLink *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* check the testexpr, if any, and the operator/keyword */
loc = leftmostLoc(exprLocation(sublink->testexpr),
sublink->location);
}
break;
case T_FieldSelect:
/* just use argument's location */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation((Node *) ((const FieldSelect *) expr)->arg);
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_FieldStore:
/* just use argument's location */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation((Node *) ((const FieldStore *) expr)->arg);
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_RelabelType:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const RelabelType *rexpr = (const RelabelType *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* Much as above */
loc = leftmostLoc(rexpr->location,
exprLocation((Node *) rexpr->arg));
}
break;
case T_CoerceViaIO:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const CoerceViaIO *cexpr = (const CoerceViaIO *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* Much as above */
loc = leftmostLoc(cexpr->location,
exprLocation((Node *) cexpr->arg));
}
break;
case T_ArrayCoerceExpr:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const ArrayCoerceExpr *cexpr = (const ArrayCoerceExpr *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* Much as above */
loc = leftmostLoc(cexpr->location,
exprLocation((Node *) cexpr->arg));
}
break;
case T_ConvertRowtypeExpr:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const ConvertRowtypeExpr *cexpr = (const ConvertRowtypeExpr *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* Much as above */
loc = leftmostLoc(cexpr->location,
exprLocation((Node *) cexpr->arg));
}
break;
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
case T_CollateExpr:
/* just use argument's location */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation((Node *) ((const CollateExpr *) expr)->arg);
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_CaseExpr:
/* CASE keyword should always be the first thing */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const CaseExpr *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_CaseWhen:
/* WHEN keyword should always be the first thing */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const CaseWhen *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_ArrayExpr:
/* the location points at ARRAY or [, which must be leftmost */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const ArrayExpr *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_RowExpr:
/* the location points at ROW or (, which must be leftmost */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const RowExpr *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_RowCompareExpr:
/* just use leftmost argument's location */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation((Node *) ((const RowCompareExpr *) expr)->largs);
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_CoalesceExpr:
/* COALESCE keyword should always be the first thing */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const CoalesceExpr *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_MinMaxExpr:
/* GREATEST/LEAST keyword should always be the first thing */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const MinMaxExpr *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
Improve parsetree representation of special functions such as CURRENT_DATE. We implement a dozen or so parameterless functions that the SQL standard defines special syntax for. Up to now, that was done by converting them into more or less ad-hoc constructs such as "'now'::text::date". That's messy for multiple reasons: it exposes what should be implementation details to users, and performance is worse than it needs to be in several cases. To improve matters, invent a new expression node type SQLValueFunction that can represent any of these parameterless functions. Bump catversion because this changes stored parsetrees for rules. Discussion: <30058.1463091294@sss.pgh.pa.us>
2016-08-16 20:33:01 -04:00
case T_SQLValueFunction:
/* function keyword should always be the first thing */
loc = ((const SQLValueFunction *) expr)->location;
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_XmlExpr:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const XmlExpr *xexpr = (const XmlExpr *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* consider both function name and leftmost arg */
loc = leftmostLoc(xexpr->location,
exprLocation((Node *) xexpr->args));
}
break;
case T_NullTest:
Add parse location fields to NullTest and BooleanTest structs. We did not need a location tag on NullTest or BooleanTest before, because no error messages referred directly to their locations. That's planned to change though, so add these fields in a separate housekeeping commit. Catversion bump because stored rules may change.
2015-02-22 14:40:27 -05:00
{
const NullTest *nexpr = (const NullTest *) expr;
/* Much as above */
loc = leftmostLoc(nexpr->location,
exprLocation((Node *) nexpr->arg));
}
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_BooleanTest:
Add parse location fields to NullTest and BooleanTest structs. We did not need a location tag on NullTest or BooleanTest before, because no error messages referred directly to their locations. That's planned to change though, so add these fields in a separate housekeeping commit. Catversion bump because stored rules may change.
2015-02-22 14:40:27 -05:00
{
const BooleanTest *bexpr = (const BooleanTest *) expr;
/* Much as above */
loc = leftmostLoc(bexpr->location,
exprLocation((Node *) bexpr->arg));
}
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_CoerceToDomain:
{
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
const CoerceToDomain *cexpr = (const CoerceToDomain *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* Much as above */
loc = leftmostLoc(cexpr->location,
exprLocation((Node *) cexpr->arg));
}
break;
case T_CoerceToDomainValue:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const CoerceToDomainValue *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_SetToDefault:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const SetToDefault *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_TargetEntry:
/* just use argument's location */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation((Node *) ((const TargetEntry *) expr)->expr);
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
Add a bunch of new error location reports to parse-analysis error messages. There are still some weak spots around JOIN USING and relation alias lists, but most errors reported within backend/parser/ now have locations.
2008-09-01 20:42:46 +00:00
case T_IntoClause:
/* use the contained RangeVar's location --- close enough */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation((Node *) ((const IntoClause *) expr)->rel);
Add a bunch of new error location reports to parse-analysis error messages. There are still some weak spots around JOIN USING and relation alias lists, but most errors reported within backend/parser/ now have locations.
2008-09-01 20:42:46 +00:00
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_List:
{
/* report location of first list member that has a location */
ListCell *lc;
loc = -1; /* just to suppress compiler warning */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
foreach(lc, (const List *) expr)
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
{
loc = exprLocation((Node *) lfirst(lc));
if (loc >= 0)
break;
}
}
break;
case T_A_Expr:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const A_Expr *aexpr = (const A_Expr *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* use leftmost of operator or left operand (if any) */
/* we assume right operand can't be to left of operator */
loc = leftmostLoc(aexpr->location,
exprLocation(aexpr->lexpr));
}
break;
case T_ColumnRef:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const ColumnRef *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_ParamRef:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const ParamRef *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_A_Const:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const A_Const *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_FuncCall:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const FuncCall *fc = (const FuncCall *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/* consider both function name and leftmost arg */
Support ORDER BY within aggregate function calls, at long last providing a non-kluge method for controlling the order in which values are fed to an aggregate function. At the same time eliminate the old implementation restriction that DISTINCT was only supported for single-argument aggregates. Possibly release-notable behavioral change: formerly, agg(DISTINCT x) dropped null values of x unconditionally. Now, it does so only if the agg transition function is strict; otherwise nulls are treated as DISTINCT normally would, ie, you get one copy. Andrew Gierth, reviewed by Hitoshi Harada
2009-12-15 17:57:48 +00:00
/* (we assume any ORDER BY nodes must be to right of name) */
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
loc = leftmostLoc(fc->location,
exprLocation((Node *) fc->args));
}
break;
case T_A_ArrayExpr:
/* the location points at ARRAY or [, which must be leftmost */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const A_ArrayExpr *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
case T_ResTarget:
/* we need not examine the contained expression (if any) */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const ResTarget *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
case T_MultiAssignRef:
loc = exprLocation(((const MultiAssignRef *) expr)->source);
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_TypeCast:
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 15:20:04 -04:00
const TypeCast *tc = (const TypeCast *) expr;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
/*
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
* This could represent CAST(), ::, or TypeName 'literal', so
* any of the components might be leftmost.
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
*/
loc = exprLocation(tc->arg);
Make backend header files C++ safe This alters various incidental uses of C++ key words to use other similar identifiers, so that a C++ compiler won't choke outright. You still (probably) need extern "C" { }; around the inclusion of backend headers. based on a patch by Kurt Harriman <harriman@acm.org> Also add a script cpluspluscheck to check for C++ compatibility in the future. As of right now, this passes without error for me.
2009-07-16 06:33:46 +00:00
loc = leftmostLoc(loc, tc->typeName->location);
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
loc = leftmostLoc(loc, tc->location);
}
break;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
case T_CollateClause:
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
/* just use argument's location */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation(((const CollateClause *) expr)->arg);
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
break;
Add a bunch of new error location reports to parse-analysis error messages. There are still some weak spots around JOIN USING and relation alias lists, but most errors reported within backend/parser/ now have locations.
2008-09-01 20:42:46 +00:00
case T_SortBy:
/* just use argument's location (ignore operator, if any) */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation(((const SortBy *) expr)->node);
Add a bunch of new error location reports to parse-analysis error messages. There are still some weak spots around JOIN USING and relation alias lists, but most errors reported within backend/parser/ now have locations.
2008-09-01 20:42:46 +00:00
break;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
case T_WindowDef:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const WindowDef *) expr)->location;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
break;
Redesign tablesample method API, and do extensive code review. The original implementation of TABLESAMPLE modeled the tablesample method API on index access methods, which wasn't a good choice because, without specialized DDL commands, there's no way to build an extension that can implement a TSM. (Raw inserts into system catalogs are not an acceptable thing to do, because we can't undo them during DROP EXTENSION, nor will pg_upgrade behave sanely.) Instead adopt an API more like procedural language handlers or foreign data wrappers, wherein the only SQL-level support object needed is a single handler function identified by having a special return type. This lets us get rid of the supporting catalog altogether, so that no custom DDL support is needed for the feature. Adjust the API so that it can support non-constant tablesample arguments (the original coding assumed we could evaluate the argument expressions at ExecInitSampleScan time, which is undesirable even if it weren't outright unsafe), and discourage sampling methods from looking at invisible tuples. Make sure that the BERNOULLI and SYSTEM methods are genuinely repeatable within and across queries, as required by the SQL standard, and deal more honestly with methods that can't support that requirement. Make a full code-review pass over the tablesample additions, and fix assorted bugs, omissions, infelicities, and cosmetic issues (such as failure to put the added code stanzas in a consistent ordering). Improve EXPLAIN's output of tablesample plans, too. Back-patch to 9.5 so that we don't have to support the original API in production.
2015-07-25 14:39:00 -04:00
case T_RangeTableSample:
loc = ((const RangeTableSample *) expr)->location;
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_TypeName:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const TypeName *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
Support multi-argument UNNEST(), and TABLE() syntax for multiple functions. This patch adds the ability to write TABLE( function1(), function2(), ...) as a single FROM-clause entry. The result is the concatenation of the first row from each function, followed by the second row from each function, etc; with NULLs inserted if any function produces fewer rows than others. This is believed to be a much more useful behavior than what Postgres currently does with multiple SRFs in a SELECT list. This syntax also provides a reasonable way to combine use of column definition lists with WITH ORDINALITY: put the column definition list inside TABLE(), where it's clear that it doesn't control the ordinality column as well. Also implement SQL-compliant multiple-argument UNNEST(), by turning UNNEST(a,b,c) into TABLE(unnest(a), unnest(b), unnest(c)). The SQL standard specifies TABLE() with only a single function, not multiple functions, and it seems to require an implicit UNNEST() which is not what this patch does. There may be something wrong with that reading of the spec, though, because if it's right then the spec's TABLE() is just a pointless alternative spelling of UNNEST(). After further review of that, we might choose to adopt a different syntax for what this patch does, but in any case this functionality seems clearly worthwhile. Andrew Gierth, reviewed by Zoltán Böszörményi and Heikki Linnakangas, and significantly revised by me
2013-11-21 19:37:02 -05:00
case T_ColumnDef:
loc = ((const ColumnDef *) expr)->location;
break;
Merge the Constraint and FkConstraint node types into a single type. This was foreseen to be a good idea long ago, but nobody had got round to doing it. The recent patch for deferred unique constraints made transformConstraintAttrs() ugly enough that I decided it was time. This change will also greatly simplify parsing of deferred CHECK constraints, if anyone ever gets around to implementing that. While at it, add a location field to Constraint, and use that to provide an error cursor for some of the constraint-related error messages.
2009-07-30 02:45:38 +00:00
case T_Constraint:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const Constraint *) expr)->location;
Merge the Constraint and FkConstraint node types into a single type. This was foreseen to be a good idea long ago, but nobody had got round to doing it. The recent patch for deferred unique constraints made transformConstraintAttrs() ugly enough that I decided it was time. This change will also greatly simplify parsing of deferred CHECK constraints, if anyone ever gets around to implementing that. While at it, add a location field to Constraint, and use that to provide an error cursor for some of the constraint-related error messages.
2009-07-30 02:45:38 +00:00
break;
Support ordered-set (WITHIN GROUP) aggregates. This patch introduces generic support for ordered-set and hypothetical-set aggregate functions, as well as implementations of the instances defined in SQL:2008 (percentile_cont(), percentile_disc(), rank(), dense_rank(), percent_rank(), cume_dist()). We also added mode() though it is not in the spec, as well as versions of percentile_cont() and percentile_disc() that can compute multiple percentile values in one pass over the data. Unlike the original submission, this patch puts full control of the sorting process in the hands of the aggregate's support functions. To allow the support functions to find out how they're supposed to sort, a new API function AggGetAggref() is added to nodeAgg.c. This allows retrieval of the aggregate call's Aggref node, which may have other uses beyond the immediate need. There is also support for ordered-set aggregates to install cleanup callback functions, so that they can be sure that infrastructure such as tuplesort objects gets cleaned up. In passing, make some fixes in the recently-added support for variadic aggregates, and make some editorial adjustments in the recent FILTER additions for aggregates. Also, simplify use of IsBinaryCoercible() by allowing it to succeed whenever the target type is ANY or ANYELEMENT. It was inconsistent that it dealt with other polymorphic target types but not these. Atri Sharma and Andrew Gierth; reviewed by Pavel Stehule and Vik Fearing, and rather heavily editorialized upon by Tom Lane
2013-12-23 16:11:35 -05:00
case T_FunctionParameter:
/* just use typename's location */
loc = exprLocation((Node *) ((const FunctionParameter *) expr)->argType);
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
case T_XmlSerialize:
/* XMLSERIALIZE keyword should always be the first thing */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const XmlSerialize *) expr)->location;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
break;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingSet:
loc = ((const GroupingSet *) expr)->location;
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_WithClause:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const WithClause *) expr)->location;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
break;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_InferClause:
loc = ((const InferClause *) expr)->location;
break;
case T_OnConflictClause:
loc = ((const OnConflictClause *) expr)->location;
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_CommonTableExpr:
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = ((const CommonTableExpr *) expr)->location;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
break;
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
case T_PlaceHolderVar:
/* just use argument's location */
Add const qualifiers to node inspection functions Thomas Munro
2011-12-07 21:46:56 +02:00
loc = exprLocation((Node *) ((const PlaceHolderVar *) expr)->phexpr);
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
break;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_InferenceElem:
/* just use nested expr's location */
loc = exprLocation((Node *) ((const InferenceElem *) expr)->expr);
break;
Code review focused on new node types added by partitioning support. Fix failure to check that we got a plain Const from const-simplification of a coercion request. This is the cause of bug #14666 from Tian Bing: there is an int4 to money cast, but it's only stable not immutable (because of dependence on lc_monetary), resulting in a FuncExpr that the code was miserably unequipped to deal with, or indeed even to notice that it was failing to deal with. Add test cases around this coercion behavior. In view of the above, sprinkle the code liberally with castNode() macros, in hope of catching the next such bug a bit sooner. Also, change some functions that were randomly declared to take Node* to take more specific pointer types. And change some struct fields that were declared Node* but could be given more specific types, allowing removal of assorted explicit casts. Place PARTITION_MAX_KEYS check a bit closer to the code it's protecting. Likewise check only-one-key-for-list-partitioning restriction in a less random place. Avoid not-per-project-style usages like !strcmp(...). Fix assorted failures to avoid scribbling on the input of parse transformation. I'm not sure how necessary this is, but it's entirely silly for these functions to be expending cycles to avoid that and not getting it right. Add guards against partitioning on system columns. Put backend/nodes/ support code into an order that matches handling of these node types elsewhere. Annotate the fact that somebody added location fields to PartitionBoundSpec and PartitionRangeDatum but forgot to handle them in outfuncs.c/readfuncs.c. This is fairly harmless for production purposes (since readfuncs.c would just substitute -1 anyway) but it's still bogus. It's not worth forcing a post-beta1 initdb just to fix this, but if we have another reason to force initdb before 10.0, we should go back and clean this up. Contrariwise, somebody added location fields to PartitionElem and PartitionSpec but forgot to teach exprLocation() about them. Consolidate duplicative code in transformPartitionBound(). Improve a couple of error messages. Improve assorted commentary. Re-pgindent the files touched by this patch; this affects a few comment blocks that must have been added quite recently. Report: https://postgr.es/m/20170524024550.29935.14396@wrigleys.postgresql.org
2017-05-28 23:20:28 -04:00
case T_PartitionElem:
loc = ((const PartitionElem *) expr)->location;
break;
case T_PartitionSpec:
loc = ((const PartitionSpec *) expr)->location;
break;
Implement table partitioning. Table partitioning is like table inheritance and reuses much of the existing infrastructure, but there are some important differences. The parent is called a partitioned table and is always empty; it may not have indexes or non-inherited constraints, since those make no sense for a relation with no data of its own. The children are called partitions and contain all of the actual data. Each partition has an implicit partitioning constraint. Multiple inheritance is not allowed, and partitioning and inheritance can't be mixed. Partitions can't have extra columns and may not allow nulls unless the parent does. Tuples inserted into the parent are automatically routed to the correct partition, so tuple-routing ON INSERT triggers are not needed. Tuple routing isn't yet supported for partitions which are foreign tables, and it doesn't handle updates that cross partition boundaries. Currently, tables can be range-partitioned or list-partitioned. List partitioning is limited to a single column, but range partitioning can involve multiple columns. A partitioning "column" can be an expression. Because table partitioning is less general than table inheritance, it is hoped that it will be easier to reason about properties of partitions, and therefore that this will serve as a better foundation for a variety of possible optimizations, including query planner optimizations. The tuple routing based which this patch does based on the implicit partitioning constraints is an example of this, but it seems likely that many other useful optimizations are also possible. Amit Langote, reviewed and tested by Robert Haas, Ashutosh Bapat, Amit Kapila, Rajkumar Raghuwanshi, Corey Huinker, Jaime Casanova, Rushabh Lathia, Erik Rijkers, among others. Minor revisions by me.
2016-12-07 13:17:43 -05:00
case T_PartitionBoundSpec:
loc = ((const PartitionBoundSpec *) expr)->location;
break;
case T_PartitionRangeDatum:
loc = ((const PartitionRangeDatum *) expr)->location;
break;
Extend the parser location infrastructure to include a location field in most node types used in expression trees (both before and after parse analysis). This allows us to place an error cursor in many situations where we formerly could not, because the information wasn't available beyond the very first level of parse analysis. There's a fair amount of work still to be done to persuade individual ereport() calls to actually include an error location, but this gets the initdb-forcing part of the work out of the way; and the situation is already markedly better than before for complaints about unimplementable implicit casts, such as CASE and UNION constructs with incompatible alternative data types. Per my proposal of a few days ago.
2008-08-28 23:09:48 +00:00
default:
/* for any other node type it's just unknown... */
loc = -1;
break;
}
return loc;
}
/*
* leftmostLoc - support for exprLocation
*
* Take the minimum of two parse location values, but ignore unknowns
*/
static int
leftmostLoc(int loc1, int loc2)
{
if (loc1 < 0)
return loc2;
else if (loc2 < 0)
return loc1;
else
return Min(loc1, loc2);
}
Refactor to reduce code duplication for function property checking. As noted by Andres Freund, we'd accumulated quite a few similar functions in clauses.c that examine all functions in an expression tree to see if they satisfy some boolean test. Reduce the duplication by inventing a function check_functions_in_node() that applies a simple callback function to each SQL function OID appearing in a given expression node. This also fixes some arguable oversights; for example, contain_mutable_functions() did not check aggregate or window functions for mutability. I doubt that that represents a live bug at the moment, because we don't really consider mutability for aggregates; but it might someday be one. I chose to put check_functions_in_node() in nodeFuncs.c because it seemed like other modules might wish to use it in future. That in turn forced moving set_opfuncid() et al into nodeFuncs.c, as the alternative was for nodeFuncs.c to depend on optimizer/setrefs.c which didn't seem very clean. In passing, teach contain_leaked_vars_walker() about a few more expression node types it can safely look through, and improve the rather messy and undercommented code in has_parallel_hazard_walker(). Discussion: <20160527185853.ziol2os2zskahl7v@alap3.anarazel.de>
2016-06-10 16:03:37 -04:00
/*
* fix_opfuncids
* Calculate opfuncid field from opno for each OpExpr node in given tree.
* The given tree can be anything expression_tree_walker handles.
*
* The argument is modified in-place. (This is OK since we'd want the
* same change for any node, even if it gets visited more than once due to
* shared structure.)
*/
void
fix_opfuncids(Node *node)
{
/* This tree walk requires no special setup, so away we go... */
fix_opfuncids_walker(node, NULL);
}
static bool
fix_opfuncids_walker(Node *node, void *context)
{
if (node == NULL)
return false;
if (IsA(node, OpExpr))
set_opfuncid((OpExpr *) node);
else if (IsA(node, DistinctExpr))
set_opfuncid((OpExpr *) node); /* rely on struct equivalence */
else if (IsA(node, NullIfExpr))
set_opfuncid((OpExpr *) node); /* rely on struct equivalence */
else if (IsA(node, ScalarArrayOpExpr))
set_sa_opfuncid((ScalarArrayOpExpr *) node);
return expression_tree_walker(node, fix_opfuncids_walker, context);
}
/*
* set_opfuncid
* Set the opfuncid (procedure OID) in an OpExpr node,
* if it hasn't been set already.
*
* Because of struct equivalence, this can also be used for
* DistinctExpr and NullIfExpr nodes.
*/
void
set_opfuncid(OpExpr *opexpr)
{
if (opexpr->opfuncid == InvalidOid)
opexpr->opfuncid = get_opcode(opexpr->opno);
}
/*
* set_sa_opfuncid
* As above, for ScalarArrayOpExpr nodes.
*/
void
set_sa_opfuncid(ScalarArrayOpExpr *opexpr)
{
if (opexpr->opfuncid == InvalidOid)
opexpr->opfuncid = get_opcode(opexpr->opno);
}
/*
* check_functions_in_node -
* apply checker() to each function OID contained in given expression node
*
Change TRUE/FALSE to true/false The lower case spellings are C and C++ standard and are used in most parts of the PostgreSQL sources. The upper case spellings are only used in some files/modules. So standardize on the standard spellings. The APIs for ICU, Perl, and Windows define their own TRUE and FALSE, so those are left as is when using those APIs. In code comments, we use the lower-case spelling for the C concepts and keep the upper-case spelling for the SQL concepts. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
2017-08-16 00:22:32 -04:00
* Returns true if the checker() function does; for nodes representing more
* than one function call, returns true if the checker() function does so
* for any of those functions. Returns false if node does not invoke any
Refactor to reduce code duplication for function property checking. As noted by Andres Freund, we'd accumulated quite a few similar functions in clauses.c that examine all functions in an expression tree to see if they satisfy some boolean test. Reduce the duplication by inventing a function check_functions_in_node() that applies a simple callback function to each SQL function OID appearing in a given expression node. This also fixes some arguable oversights; for example, contain_mutable_functions() did not check aggregate or window functions for mutability. I doubt that that represents a live bug at the moment, because we don't really consider mutability for aggregates; but it might someday be one. I chose to put check_functions_in_node() in nodeFuncs.c because it seemed like other modules might wish to use it in future. That in turn forced moving set_opfuncid() et al into nodeFuncs.c, as the alternative was for nodeFuncs.c to depend on optimizer/setrefs.c which didn't seem very clean. In passing, teach contain_leaked_vars_walker() about a few more expression node types it can safely look through, and improve the rather messy and undercommented code in has_parallel_hazard_walker(). Discussion: <20160527185853.ziol2os2zskahl7v@alap3.anarazel.de>
2016-06-10 16:03:37 -04:00
* SQL-visible function. Caller must not pass node == NULL.
*
* This function examines only the given node; it does not recurse into any
* sub-expressions. Callers typically prefer to keep control of the recursion
* for themselves, in case additional checks should be made, or because they
* have special rules about which parts of the tree need to be visited.
*
Code review for NextValueExpr expression node type. Add missing infrastructure for this node type, notably in ruleutils.c where its lack could demonstrably cause EXPLAIN to fail. Add outfuncs/readfuncs support. (outfuncs support is useful today for debugging purposes. The readfuncs support may never be needed, since at present it would only matter for parallel query and NextValueExpr should never appear in a parallelizable query; but it seems like a bad idea to have a primnode type that isn't fully supported here.) Teach planner infrastructure that NextValueExpr is a volatile, parallel-unsafe, non-leaky expression node with cost cpu_operator_cost. Given its limited scope of usage, there *might* be no live bug today from the lack of that knowledge, but it's certainly going to bite us on the rear someday. Teach pg_stat_statements about the new node type, too. While at it, also teach cost_qual_eval() that MinMaxExpr, SQLValueFunction, XmlExpr, and CoerceToDomain should be charged as cpu_operator_cost. Failing to do this for SQLValueFunction was an oversight in my commit 0bb51aa96. The others are longer-standing oversights, but no time like the present to fix them. (In principle, CoerceToDomain could have cost much higher than this, but it doesn't presently seem worth trying to examine the domain's constraints here.) Modify execExprInterp.c to execute NextValueExpr as an out-of-line function; it seems quite unlikely to me that it's worth insisting that it be inlined in all expression eval methods. Besides, providing the out-of-line function doesn't stop anyone from inlining if they want to. Adjust some places where NextValueExpr support had been inserted with the aid of a dartboard rather than keeping it in the same order as elsewhere. Discussion: https://postgr.es/m/23862.1499981661@sss.pgh.pa.us
2017-07-14 15:25:43 -04:00
* Note: we ignore MinMaxExpr, SQLValueFunction, XmlExpr, CoerceToDomain,
* and NextValueExpr nodes, because they do not contain SQL function OIDs.
* However, they can invoke SQL-visible functions, so callers should take
* thought about how to treat them.
Refactor to reduce code duplication for function property checking. As noted by Andres Freund, we'd accumulated quite a few similar functions in clauses.c that examine all functions in an expression tree to see if they satisfy some boolean test. Reduce the duplication by inventing a function check_functions_in_node() that applies a simple callback function to each SQL function OID appearing in a given expression node. This also fixes some arguable oversights; for example, contain_mutable_functions() did not check aggregate or window functions for mutability. I doubt that that represents a live bug at the moment, because we don't really consider mutability for aggregates; but it might someday be one. I chose to put check_functions_in_node() in nodeFuncs.c because it seemed like other modules might wish to use it in future. That in turn forced moving set_opfuncid() et al into nodeFuncs.c, as the alternative was for nodeFuncs.c to depend on optimizer/setrefs.c which didn't seem very clean. In passing, teach contain_leaked_vars_walker() about a few more expression node types it can safely look through, and improve the rather messy and undercommented code in has_parallel_hazard_walker(). Discussion: <20160527185853.ziol2os2zskahl7v@alap3.anarazel.de>
2016-06-10 16:03:37 -04:00
*/
bool
check_functions_in_node(Node *node, check_function_callback checker,
void *context)
{
switch (nodeTag(node))
{
case T_Aggref:
{
Aggref *expr = (Aggref *) node;
if (checker(expr->aggfnoid, context))
return true;
}
break;
case T_WindowFunc:
{
WindowFunc *expr = (WindowFunc *) node;
if (checker(expr->winfnoid, context))
return true;
}
break;
case T_FuncExpr:
{
FuncExpr *expr = (FuncExpr *) node;
if (checker(expr->funcid, context))
return true;
}
break;
case T_OpExpr:
case T_DistinctExpr: /* struct-equivalent to OpExpr */
case T_NullIfExpr: /* struct-equivalent to OpExpr */
{
OpExpr *expr = (OpExpr *) node;
/* Set opfuncid if it wasn't set already */
set_opfuncid(expr);
if (checker(expr->opfuncid, context))
return true;
}
break;
case T_ScalarArrayOpExpr:
{
ScalarArrayOpExpr *expr = (ScalarArrayOpExpr *) node;
set_sa_opfuncid(expr);
if (checker(expr->opfuncid, context))
return true;
}
break;
case T_CoerceViaIO:
{
CoerceViaIO *expr = (CoerceViaIO *) node;
Oid iofunc;
Oid typioparam;
bool typisvarlena;
/* check the result type's input function */
getTypeInputInfo(expr->resulttype,
&iofunc, &typioparam);
if (checker(iofunc, context))
return true;
/* check the input type's output function */
getTypeOutputInfo(exprType((Node *) expr->arg),
&iofunc, &typisvarlena);
if (checker(iofunc, context))
return true;
}
break;
case T_RowCompareExpr:
{
RowCompareExpr *rcexpr = (RowCompareExpr *) node;
ListCell *opid;
foreach(opid, rcexpr->opnos)
{
Oid opfuncid = get_opcode(lfirst_oid(opid));
if (checker(opfuncid, context))
return true;
}
}
break;
default:
break;
}
return false;
}
Massive commit to run PGINDENT on all *.c and *.h files.
1997-09-07 05:04:48 +00:00
/*
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
* Standard expression-tree walking support
*
* We used to have near-duplicate code in many different routines that
* understood how to recurse through an expression node tree. That was
* a pain to maintain, and we frequently had bugs due to some particular
* routine neglecting to support a particular node type. In most cases,
* these routines only actually care about certain node types, and don't
* care about other types except insofar as they have to recurse through
* non-primitive node types. Therefore, we now provide generic tree-walking
* logic to consolidate the redundant "boilerplate" code. There are
* two versions: expression_tree_walker() and expression_tree_mutator().
*/
/*
* expression_tree_walker() is designed to support routines that traverse
* a tree in a read-only fashion (although it will also work for routines
* that modify nodes in-place but never add/delete/replace nodes).
* A walker routine should look like this:
*
* bool my_walker (Node *node, my_struct *context)
* {
* if (node == NULL)
* return false;
* // check for nodes that special work is required for, eg:
* if (IsA(node, Var))
* {
* ... do special actions for Var nodes
* }
* else if (IsA(node, ...))
* {
* ... do special actions for other node types
* }
* // for any node type not specially processed, do:
* return expression_tree_walker(node, my_walker, (void *) context);
* }
*
* The "context" argument points to a struct that holds whatever context
* information the walker routine needs --- it can be used to return data
* gathered by the walker, too. This argument is not touched by
* expression_tree_walker, but it is passed down to recursive sub-invocations
* of my_walker. The tree walk is started from a setup routine that
* fills in the appropriate context struct, calls my_walker with the top-level
* node of the tree, and then examines the results.
*
* The walker routine should return "false" to continue the tree walk, or
* "true" to abort the walk and immediately return "true" to the top-level
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 12:12:18 -04:00
* caller. This can be used to short-circuit the traversal if the walker
* has found what it came for. "false" is returned to the top-level caller
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
* iff no invocation of the walker returned "true".
*
* The node types handled by expression_tree_walker include all those
* normally found in target lists and qualifier clauses during the planning
* stage. In particular, it handles List nodes since a cnf-ified qual clause
* will have List structure at the top level, and it handles TargetEntry nodes
* so that a scan of a target list can be handled without additional code.
* Also, RangeTblRef, FromExpr, JoinExpr, and SetOperationStmt nodes are
* handled, so that query jointrees and setOperation trees can be processed
* without additional code.
*
* expression_tree_walker will handle SubLink nodes by recursing normally
* into the "testexpr" subtree (which is an expression belonging to the outer
* plan). It will also call the walker on the sub-Query node; however, when
* expression_tree_walker itself is called on a Query node, it does nothing
* and returns "false". The net effect is that unless the walker does
* something special at a Query node, sub-selects will not be visited during
* an expression tree walk. This is exactly the behavior wanted in many cases
* --- and for those walkers that do want to recurse into sub-selects, special
* behavior is typically needed anyway at the entry to a sub-select (such as
* incrementing a depth counter). A walker that wants to examine sub-selects
* should include code along the lines of:
*
* if (IsA(node, Query))
* {
* adjust context for subquery;
* result = query_tree_walker((Query *) node, my_walker, context,
* 0); // adjust flags as needed
* restore context if needed;
* return result;
* }
*
* query_tree_walker is a convenience routine (see below) that calls the
* walker on all the expression subtrees of the given Query node.
*
* expression_tree_walker will handle SubPlan nodes by recursing normally
* into the "testexpr" and the "args" list (which are expressions belonging to
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 12:12:18 -04:00
* the outer plan). It will not touch the completed subplan, however. Since
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
* there is no link to the original Query, it is not possible to recurse into
* subselects of an already-planned expression tree. This is OK for current
* uses, but may need to be revisited in future.
*/
bool
expression_tree_walker(Node *node,
bool (*walker) (),
void *context)
{
ListCell *temp;
/*
* The walker has already visited the current node, and so we need only
* recurse into any sub-nodes it has.
*
* We assume that the walker is not interested in List nodes per se, so
* when we expect a List we just recurse directly to self without
* bothering to call the walker.
*/
if (node == NULL)
return false;
/* Guard against stack overflow due to overly complex expressions */
check_stack_depth();
switch (nodeTag(node))
{
case T_Var:
case T_Const:
case T_Param:
case T_CaseTestExpr:
Code review for NextValueExpr expression node type. Add missing infrastructure for this node type, notably in ruleutils.c where its lack could demonstrably cause EXPLAIN to fail. Add outfuncs/readfuncs support. (outfuncs support is useful today for debugging purposes. The readfuncs support may never be needed, since at present it would only matter for parallel query and NextValueExpr should never appear in a parallelizable query; but it seems like a bad idea to have a primnode type that isn't fully supported here.) Teach planner infrastructure that NextValueExpr is a volatile, parallel-unsafe, non-leaky expression node with cost cpu_operator_cost. Given its limited scope of usage, there *might* be no live bug today from the lack of that knowledge, but it's certainly going to bite us on the rear someday. Teach pg_stat_statements about the new node type, too. While at it, also teach cost_qual_eval() that MinMaxExpr, SQLValueFunction, XmlExpr, and CoerceToDomain should be charged as cpu_operator_cost. Failing to do this for SQLValueFunction was an oversight in my commit 0bb51aa96. The others are longer-standing oversights, but no time like the present to fix them. (In principle, CoerceToDomain could have cost much higher than this, but it doesn't presently seem worth trying to examine the domain's constraints here.) Modify execExprInterp.c to execute NextValueExpr as an out-of-line function; it seems quite unlikely to me that it's worth insisting that it be inlined in all expression eval methods. Besides, providing the out-of-line function doesn't stop anyone from inlining if they want to. Adjust some places where NextValueExpr support had been inserted with the aid of a dartboard rather than keeping it in the same order as elsewhere. Discussion: https://postgr.es/m/23862.1499981661@sss.pgh.pa.us
2017-07-14 15:25:43 -04:00
case T_SQLValueFunction:
case T_CoerceToDomainValue:
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_SetToDefault:
case T_CurrentOfExpr:
Identity columns This is the SQL standard-conforming variant of PostgreSQL's serial columns. It fixes a few usability issues that serial columns have: - CREATE TABLE / LIKE copies default but refers to same sequence - cannot add/drop serialness with ALTER TABLE - dropping default does not drop sequence - need to grant separate privileges to sequence - other slight weirdnesses because serial is some kind of special macro Reviewed-by: Vitaly Burovoy <vitaly.burovoy@gmail.com>
2017-04-06 08:33:16 -04:00
case T_NextValueExpr:
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_RangeTblRef:
Support ORDER BY within aggregate function calls, at long last providing a non-kluge method for controlling the order in which values are fed to an aggregate function. At the same time eliminate the old implementation restriction that DISTINCT was only supported for single-argument aggregates. Possibly release-notable behavioral change: formerly, agg(DISTINCT x) dropped null values of x unconditionally. Now, it does so only if the agg transition function is strict; otherwise nulls are treated as DISTINCT normally would, ie, you get one copy. Andrew Gierth, reviewed by Hitoshi Harada
2009-12-15 17:57:48 +00:00
case T_SortGroupClause:
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
/* primitive node types with no expression subnodes */
break;
WITH CHECK OPTION support for auto-updatable VIEWs For simple views which are automatically updatable, this patch allows the user to specify what level of checking should be done on records being inserted or updated. For 'LOCAL CHECK', new tuples are validated against the conditionals of the view they are being inserted into, while for 'CASCADED CHECK' the new tuples are validated against the conditionals for all views involved (from the top down). This option is part of the SQL specification. Dean Rasheed, reviewed by Pavel Stehule
2013-07-18 17:10:16 -04:00
case T_WithCheckOption:
return walker(((WithCheckOption *) node)->qual, context);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_Aggref:
{
Aggref *expr = (Aggref *) node;
/* recurse directly on List */
Support ordered-set (WITHIN GROUP) aggregates. This patch introduces generic support for ordered-set and hypothetical-set aggregate functions, as well as implementations of the instances defined in SQL:2008 (percentile_cont(), percentile_disc(), rank(), dense_rank(), percent_rank(), cume_dist()). We also added mode() though it is not in the spec, as well as versions of percentile_cont() and percentile_disc() that can compute multiple percentile values in one pass over the data. Unlike the original submission, this patch puts full control of the sorting process in the hands of the aggregate's support functions. To allow the support functions to find out how they're supposed to sort, a new API function AggGetAggref() is added to nodeAgg.c. This allows retrieval of the aggregate call's Aggref node, which may have other uses beyond the immediate need. There is also support for ordered-set aggregates to install cleanup callback functions, so that they can be sure that infrastructure such as tuplesort objects gets cleaned up. In passing, make some fixes in the recently-added support for variadic aggregates, and make some editorial adjustments in the recent FILTER additions for aggregates. Also, simplify use of IsBinaryCoercible() by allowing it to succeed whenever the target type is ANY or ANYELEMENT. It was inconsistent that it dealt with other polymorphic target types but not these. Atri Sharma and Andrew Gierth; reviewed by Pavel Stehule and Vik Fearing, and rather heavily editorialized upon by Tom Lane
2013-12-23 16:11:35 -05:00
if (expression_tree_walker((Node *) expr->aggdirectargs,
walker, context))
return true;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (expression_tree_walker((Node *) expr->args,
walker, context))
return true;
Support ORDER BY within aggregate function calls, at long last providing a non-kluge method for controlling the order in which values are fed to an aggregate function. At the same time eliminate the old implementation restriction that DISTINCT was only supported for single-argument aggregates. Possibly release-notable behavioral change: formerly, agg(DISTINCT x) dropped null values of x unconditionally. Now, it does so only if the agg transition function is strict; otherwise nulls are treated as DISTINCT normally would, ie, you get one copy. Andrew Gierth, reviewed by Hitoshi Harada
2009-12-15 17:57:48 +00:00
if (expression_tree_walker((Node *) expr->aggorder,
walker, context))
return true;
if (expression_tree_walker((Node *) expr->aggdistinct,
walker, context))
return true;
Implement the FILTER clause for aggregate function calls. This is SQL-standard with a few extensions, namely support for subqueries and outer references in clause expressions. catversion bump due to change in Aggref and WindowFunc. David Fetter, reviewed by Dean Rasheed.
2013-07-16 20:15:36 -04:00
if (walker((Node *) expr->aggfilter, context))
return true;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
}
break;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingFunc:
{
GroupingFunc *grouping = (GroupingFunc *) node;
if (expression_tree_walker((Node *) grouping->args,
walker, context))
return true;
}
break;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
case T_WindowFunc:
{
WindowFunc *expr = (WindowFunc *) node;
/* recurse directly on List */
if (expression_tree_walker((Node *) expr->args,
walker, context))
return true;
Implement the FILTER clause for aggregate function calls. This is SQL-standard with a few extensions, namely support for subqueries and outer references in clause expressions. catversion bump due to change in Aggref and WindowFunc. David Fetter, reviewed by Dean Rasheed.
2013-07-16 20:15:36 -04:00
if (walker((Node *) expr->aggfilter, context))
return true;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ArrayRef:
{
ArrayRef *aref = (ArrayRef *) node;
/* recurse directly for upper/lower array index lists */
if (expression_tree_walker((Node *) aref->refupperindexpr,
walker, context))
return true;
if (expression_tree_walker((Node *) aref->reflowerindexpr,
walker, context))
return true;
/* walker must see the refexpr and refassgnexpr, however */
if (walker(aref->refexpr, context))
return true;
if (walker(aref->refassgnexpr, context))
return true;
}
break;
case T_FuncExpr:
{
FuncExpr *expr = (FuncExpr *) node;
if (expression_tree_walker((Node *) expr->args,
walker, context))
return true;
}
break;
Support use of function argument names to identify which actual arguments match which function parameters. The syntax uses AS, for example funcname(value AS arg1, anothervalue AS arg2) Pavel Stehule
2009-10-08 02:39:25 +00:00
case T_NamedArgExpr:
return walker(((NamedArgExpr *) node)->arg, context);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_OpExpr:
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_DistinctExpr: /* struct-equivalent to OpExpr */
case T_NullIfExpr: /* struct-equivalent to OpExpr */
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
{
OpExpr *expr = (OpExpr *) node;
if (expression_tree_walker((Node *) expr->args,
walker, context))
return true;
}
break;
case T_ScalarArrayOpExpr:
{
ScalarArrayOpExpr *expr = (ScalarArrayOpExpr *) node;
if (expression_tree_walker((Node *) expr->args,
walker, context))
return true;
}
break;
case T_BoolExpr:
{
BoolExpr *expr = (BoolExpr *) node;
if (expression_tree_walker((Node *) expr->args,
walker, context))
return true;
}
break;
case T_SubLink:
{
SubLink *sublink = (SubLink *) node;
if (walker(sublink->testexpr, context))
return true;
/*
* Also invoke the walker on the sublink's Query node, so it
* can recurse into the sub-query if it wants to.
*/
return walker(sublink->subselect, context);
}
break;
case T_SubPlan:
{
SubPlan *subplan = (SubPlan *) node;
/* recurse into the testexpr, but not into the Plan */
if (walker(subplan->testexpr, context))
return true;
/* also examine args list */
if (expression_tree_walker((Node *) subplan->args,
walker, context))
return true;
}
break;
case T_AlternativeSubPlan:
return walker(((AlternativeSubPlan *) node)->subplans, context);
case T_FieldSelect:
return walker(((FieldSelect *) node)->arg, context);
case T_FieldStore:
{
FieldStore *fstore = (FieldStore *) node;
if (walker(fstore->arg, context))
return true;
if (walker(fstore->newvals, context))
return true;
}
break;
case T_RelabelType:
return walker(((RelabelType *) node)->arg, context);
case T_CoerceViaIO:
return walker(((CoerceViaIO *) node)->arg, context);
case T_ArrayCoerceExpr:
Support arrays over domains. Allowing arrays with a domain type as their element type was left un-done in the original domain patch, but not for any very good reason. This omission leads to such surprising results as array_agg() not working on a domain column, because the parser can't identify a suitable output type for the polymorphic aggregate. In order to fix this, first clean up the APIs of coerce_to_domain() and some internal functions in parse_coerce.c so that we consistently pass around a CoercionContext along with CoercionForm. Previously, we sometimes passed an "isExplicit" boolean flag instead, which is strictly less information; and coerce_to_domain() didn't even get that, but instead had to reverse-engineer isExplicit from CoercionForm. That's contrary to the documentation in primnodes.h that says that CoercionForm only affects display and not semantics. I don't think this change fixes any live bugs, but it makes things more consistent. The main reason for doing it though is that now build_coercion_expression() receives ccontext, which it needs in order to be able to recursively invoke coerce_to_target_type(). Next, reimplement ArrayCoerceExpr so that the node does not directly know any details of what has to be done to the individual array elements while performing the array coercion. Instead, the per-element processing is represented by a sub-expression whose input is a source array element and whose output is a target array element. This simplifies life in parse_coerce.c, because it can build that sub-expression by a recursive invocation of coerce_to_target_type(). The executor now handles the per-element processing as a compiled expression instead of hard-wired code. The main advantage of this is that we can use a single ArrayCoerceExpr to handle as many as three successive steps per element: base type conversion, typmod coercion, and domain constraint checking. The old code used two stacked ArrayCoerceExprs to handle type + typmod coercion, which was pretty inefficient, and adding yet another array deconstruction to do domain constraint checking seemed very unappetizing. In the case where we just need a single, very simple coercion function, doing this straightforwardly leads to a noticeable increase in the per-array-element runtime cost. Hence, add an additional shortcut evalfunc in execExprInterp.c that skips unnecessary overhead for that specific form of expression. The runtime speed of simple cases is within 1% or so of where it was before, while cases that previously required two levels of array processing are significantly faster. Finally, create an implicit array type for every domain type, as we do for base types, enums, etc. Everything except the array-coercion case seems to just work without further effort. Tom Lane, reviewed by Andrew Dunstan Discussion: https://postgr.es/m/9852.1499791473@sss.pgh.pa.us
2017-09-30 13:40:56 -04:00
{
ArrayCoerceExpr *acoerce = (ArrayCoerceExpr *) node;
if (walker(acoerce->arg, context))
return true;
if (walker(acoerce->elemexpr, context))
return true;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ConvertRowtypeExpr:
return walker(((ConvertRowtypeExpr *) node)->arg, context);
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
case T_CollateExpr:
return walker(((CollateExpr *) node)->arg, context);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_CaseExpr:
{
CaseExpr *caseexpr = (CaseExpr *) node;
if (walker(caseexpr->arg, context))
return true;
/* we assume walker doesn't care about CaseWhens, either */
foreach(temp, caseexpr->args)
{
Improve castNode notation by introducing list-extraction-specific variants. This extends the castNode() notation introduced by commit 5bcab1114 to provide, in one step, extraction of a list cell's pointer and coercion to a concrete node type. For example, "lfirst_node(Foo, lc)" is the same as "castNode(Foo, lfirst(lc))". Almost half of the uses of castNode that have appeared so far include a list extraction call, so this is pretty widely useful, and it saves a few more keystrokes compared to the old way. As with the previous patch, back-patch the addition of these macros to pg_list.h, so that the notation will be available when back-patching. Patch by me, after an idea of Andrew Gierth's. Discussion: https://postgr.es/m/14197.1491841216@sss.pgh.pa.us
2017-04-10 13:51:29 -04:00
CaseWhen *when = lfirst_node(CaseWhen, temp);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (walker(when->expr, context))
return true;
if (walker(when->result, context))
return true;
}
if (walker(caseexpr->defresult, context))
return true;
}
break;
case T_ArrayExpr:
return walker(((ArrayExpr *) node)->elements, context);
case T_RowExpr:
When expanding a whole-row Var into a RowExpr during ResolveNew(), attach the column alias names of the RTE referenced by the Var to the RowExpr. This is needed to allow ruleutils.c to correctly deparse FieldSelect nodes referencing such a construct. Per my recent bug report. Adding a field to RowExpr forces initdb (because of stored rules changes) so this solution is not back-patchable; which is unfortunate because 8.2 and 8.3 have this issue. But it only affects EXPLAIN for some pretty odd corner cases, so we can probably live without a solution for the back branches.
2008-10-06 17:39:26 +00:00
/* Assume colnames isn't interesting */
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
return walker(((RowExpr *) node)->args, context);
case T_RowCompareExpr:
{
RowCompareExpr *rcexpr = (RowCompareExpr *) node;
if (walker(rcexpr->largs, context))
return true;
if (walker(rcexpr->rargs, context))
return true;
}
break;
case T_CoalesceExpr:
return walker(((CoalesceExpr *) node)->args, context);
case T_MinMaxExpr:
return walker(((MinMaxExpr *) node)->args, context);
case T_XmlExpr:
{
XmlExpr *xexpr = (XmlExpr *) node;
if (walker(xexpr->named_args, context))
return true;
/* we assume walker doesn't care about arg_names */
if (walker(xexpr->args, context))
return true;
}
break;
case T_NullTest:
return walker(((NullTest *) node)->arg, context);
case T_BooleanTest:
return walker(((BooleanTest *) node)->arg, context);
case T_CoerceToDomain:
return walker(((CoerceToDomain *) node)->arg, context);
case T_TargetEntry:
return walker(((TargetEntry *) node)->expr, context);
case T_Query:
/* Do nothing with a sub-Query, per discussion above */
break;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
case T_WindowClause:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
WindowClause *wc = (WindowClause *) node;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
if (walker(wc->partitionClause, context))
return true;
if (walker(wc->orderClause, context))
return true;
Extend the set of frame options supported for window functions. This patch allows the frame to start from CURRENT ROW (in either RANGE or ROWS mode), and it also adds support for ROWS n PRECEDING and ROWS n FOLLOWING start and end points. (RANGE value PRECEDING/FOLLOWING isn't there yet --- the grammar works, but that's all.) Hitoshi Harada, reviewed by Pavel Stehule
2010-02-12 17:33:21 +00:00
if (walker(wc->startOffset, context))
return true;
if (walker(wc->endOffset, context))
return true;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
}
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_CommonTableExpr:
{
CommonTableExpr *cte = (CommonTableExpr *) node;
/*
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
* Invoke the walker on the CTE's Query node, so it can
* recurse into the sub-query if it wants to.
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
*/
return walker(cte->ctequery, context);
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_List:
foreach(temp, (List *) node)
{
if (walker((Node *) lfirst(temp), context))
return true;
}
break;
case T_FromExpr:
{
FromExpr *from = (FromExpr *) node;
if (walker(from->fromlist, context))
return true;
if (walker(from->quals, context))
return true;
}
break;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_OnConflictExpr:
{
pgindent run for 9.5
2015-05-23 21:35:49 -04:00
OnConflictExpr *onconflict = (OnConflictExpr *) node;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
if (walker((Node *) onconflict->arbiterElems, context))
return true;
if (walker(onconflict->arbiterWhere, context))
return true;
if (walker(onconflict->onConflictSet, context))
return true;
if (walker(onconflict->onConflictWhere, context))
return true;
Fix ON CONFLICT bugs that manifest when used in rules. Specifically the tlist and rti of the pseudo "excluded" relation weren't properly treated by expression_tree_walker, which lead to errors when excluded was referenced inside a rule because the varnos where not properly adjusted. Similar omissions in OffsetVarNodes and expression_tree_mutator had less impact, but should obviously be fixed nonetheless. A couple tests of for ON CONFLICT UPDATE into INSERT rule bearing relations have been added. In passing I updated a couple comments.
2015-05-13 00:13:22 +02:00
if (walker(onconflict->exclRelTlist, context))
return true;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
}
break;
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
case T_MergeAction:
{
MergeAction *action = (MergeAction *) node;
if (walker(action->targetList, context))
return true;
if (walker(action->qual, context))
return true;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_JoinExpr:
{
JoinExpr *join = (JoinExpr *) node;
if (walker(join->larg, context))
return true;
if (walker(join->rarg, context))
return true;
if (walker(join->quals, context))
return true;
/*
* alias clause, using list are deemed uninteresting.
*/
}
break;
case T_SetOperationStmt:
{
SetOperationStmt *setop = (SetOperationStmt *) node;
if (walker(setop->larg, context))
return true;
if (walker(setop->rarg, context))
return true;
/* groupClauses are deemed uninteresting */
}
break;
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
case T_PlaceHolderVar:
return walker(((PlaceHolderVar *) node)->phexpr, context);
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_InferenceElem:
return walker(((InferenceElem *) node)->expr, context);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_AppendRelInfo:
{
AppendRelInfo *appinfo = (AppendRelInfo *) node;
if (expression_tree_walker((Node *) appinfo->translated_vars,
walker, context))
return true;
}
break;
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
case T_PlaceHolderInfo:
return walker(((PlaceHolderInfo *) node)->ph_var, context);
Support multi-argument UNNEST(), and TABLE() syntax for multiple functions. This patch adds the ability to write TABLE( function1(), function2(), ...) as a single FROM-clause entry. The result is the concatenation of the first row from each function, followed by the second row from each function, etc; with NULLs inserted if any function produces fewer rows than others. This is believed to be a much more useful behavior than what Postgres currently does with multiple SRFs in a SELECT list. This syntax also provides a reasonable way to combine use of column definition lists with WITH ORDINALITY: put the column definition list inside TABLE(), where it's clear that it doesn't control the ordinality column as well. Also implement SQL-compliant multiple-argument UNNEST(), by turning UNNEST(a,b,c) into TABLE(unnest(a), unnest(b), unnest(c)). The SQL standard specifies TABLE() with only a single function, not multiple functions, and it seems to require an implicit UNNEST() which is not what this patch does. There may be something wrong with that reading of the spec, though, because if it's right then the spec's TABLE() is just a pointless alternative spelling of UNNEST(). After further review of that, we might choose to adopt a different syntax for what this patch does, but in any case this functionality seems clearly worthwhile. Andrew Gierth, reviewed by Zoltán Böszörményi and Heikki Linnakangas, and significantly revised by me
2013-11-21 19:37:02 -05:00
case T_RangeTblFunction:
return walker(((RangeTblFunction *) node)->funcexpr, context);
Redesign tablesample method API, and do extensive code review. The original implementation of TABLESAMPLE modeled the tablesample method API on index access methods, which wasn't a good choice because, without specialized DDL commands, there's no way to build an extension that can implement a TSM. (Raw inserts into system catalogs are not an acceptable thing to do, because we can't undo them during DROP EXTENSION, nor will pg_upgrade behave sanely.) Instead adopt an API more like procedural language handlers or foreign data wrappers, wherein the only SQL-level support object needed is a single handler function identified by having a special return type. This lets us get rid of the supporting catalog altogether, so that no custom DDL support is needed for the feature. Adjust the API so that it can support non-constant tablesample arguments (the original coding assumed we could evaluate the argument expressions at ExecInitSampleScan time, which is undesirable even if it weren't outright unsafe), and discourage sampling methods from looking at invisible tuples. Make sure that the BERNOULLI and SYSTEM methods are genuinely repeatable within and across queries, as required by the SQL standard, and deal more honestly with methods that can't support that requirement. Make a full code-review pass over the tablesample additions, and fix assorted bugs, omissions, infelicities, and cosmetic issues (such as failure to put the added code stanzas in a consistent ordering). Improve EXPLAIN's output of tablesample plans, too. Back-patch to 9.5 so that we don't have to support the original API in production.
2015-07-25 14:39:00 -04:00
case T_TableSampleClause:
{
TableSampleClause *tsc = (TableSampleClause *) node;
if (expression_tree_walker((Node *) tsc->args,
walker, context))
return true;
if (walker((Node *) tsc->repeatable, context))
return true;
}
break;
Support XMLTABLE query expression XMLTABLE is defined by the SQL/XML standard as a feature that allows turning XML-formatted data into relational form, so that it can be used as a <table primary> in the FROM clause of a query. This new construct provides significant simplicity and performance benefit for XML data processing; what in a client-side custom implementation was reported to take 20 minutes can be executed in 400ms using XMLTABLE. (The same functionality was said to take 10 seconds using nested PostgreSQL XPath function calls, and 5 seconds using XMLReader under PL/Python). The implemented syntax deviates slightly from what the standard requires. First, the standard indicates that the PASSING clause is optional and that multiple XML input documents may be given to it; we make it mandatory and accept a single document only. Second, we don't currently support a default namespace to be specified. This implementation relies on a new executor node based on a hardcoded method table. (Because the grammar is fixed, there is no extensibility in the current approach; further constructs can be implemented on top of this such as JSON_TABLE, but they require changes to core code.) Author: Pavel Stehule, Álvaro Herrera Extensively reviewed by: Craig Ringer Discussion: https://postgr.es/m/CAFj8pRAgfzMD-LoSmnMGybD0WsEznLHWap8DO79+-GTRAPR4qA@mail.gmail.com
2017-03-08 12:39:37 -03:00
case T_TableFunc:
{
TableFunc *tf = (TableFunc *) node;
if (walker(tf->ns_uris, context))
return true;
if (walker(tf->docexpr, context))
return true;
if (walker(tf->rowexpr, context))
return true;
if (walker(tf->colexprs, context))
return true;
if (walker(tf->coldefexprs, context))
return true;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
default:
elog(ERROR, "unrecognized node type: %d",
(int) nodeTag(node));
break;
}
return false;
}
/*
* query_tree_walker --- initiate a walk of a Query's expressions
*
* This routine exists just to reduce the number of places that need to know
* where all the expression subtrees of a Query are. Note it can be used
* for starting a walk at top level of a Query regardless of whether the
* walker intends to descend into subqueries. It is also useful for
* descending into subqueries within a walker.
Massive commit to run PGINDENT on all *.c and *.h files.
1997-09-07 05:04:48 +00:00
*
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
* Some callers want to suppress visitation of certain items in the sub-Query,
* typically because they need to process them specially, or don't actually
* want to recurse into subqueries. This is supported by the flags argument,
* which is the bitwise OR of flag values to suppress visitation of
* indicated items. (More flag bits may be added as needed.)
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
*/
bool
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
query_tree_walker(Query *query,
bool (*walker) (),
void *context,
int flags)
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
{
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
Assert(query != NULL && IsA(query, Query));
if (walker((Node *) query->targetList, context))
return true;
WITH CHECK OPTION support for auto-updatable VIEWs For simple views which are automatically updatable, this patch allows the user to specify what level of checking should be done on records being inserted or updated. For 'LOCAL CHECK', new tuples are validated against the conditionals of the view they are being inserted into, while for 'CASCADED CHECK' the new tuples are validated against the conditionals for all views involved (from the top down). This option is part of the SQL specification. Dean Rasheed, reviewed by Pavel Stehule
2013-07-18 17:10:16 -04:00
if (walker((Node *) query->withCheckOptions, context))
return true;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
if (walker((Node *) query->onConflict, context))
return true;
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
if (walker((Node *) query->mergeSourceTargetList, context))
return true;
if (walker((Node *) query->mergeActionList, context))
return true;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (walker((Node *) query->returningList, context))
return true;
if (walker((Node *) query->jointree, context))
return true;
if (walker(query->setOperations, context))
return true;
if (walker(query->havingQual, context))
return true;
if (walker(query->limitOffset, context))
return true;
if (walker(query->limitCount, context))
return true;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (!(flags & QTW_IGNORE_CTE_SUBQUERIES))
{
if (walker((Node *) query->cteList, context))
return true;
}
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
if (!(flags & QTW_IGNORE_RANGE_TABLE))
{
if (range_table_walker(query->rtable, walker, context, flags))
return true;
}
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
return false;
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
}
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
/*
* range_table_walker is just the part of query_tree_walker that scans
* a query's rangetable. This is split out since it can be useful on
* its own.
*/
bool
range_table_walker(List *rtable,
bool (*walker) (),
void *context,
int flags)
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
{
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
ListCell *rt;
foreach(rt, rtable)
{
RangeTblEntry *rte = (RangeTblEntry *) lfirst(rt);
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
/* For historical reasons, visiting RTEs is not the default */
if (flags & QTW_EXAMINE_RTES)
if (walker(rte, context))
return true;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
switch (rte->rtekind)
{
case RTE_RELATION:
Redesign tablesample method API, and do extensive code review. The original implementation of TABLESAMPLE modeled the tablesample method API on index access methods, which wasn't a good choice because, without specialized DDL commands, there's no way to build an extension that can implement a TSM. (Raw inserts into system catalogs are not an acceptable thing to do, because we can't undo them during DROP EXTENSION, nor will pg_upgrade behave sanely.) Instead adopt an API more like procedural language handlers or foreign data wrappers, wherein the only SQL-level support object needed is a single handler function identified by having a special return type. This lets us get rid of the supporting catalog altogether, so that no custom DDL support is needed for the feature. Adjust the API so that it can support non-constant tablesample arguments (the original coding assumed we could evaluate the argument expressions at ExecInitSampleScan time, which is undesirable even if it weren't outright unsafe), and discourage sampling methods from looking at invisible tuples. Make sure that the BERNOULLI and SYSTEM methods are genuinely repeatable within and across queries, as required by the SQL standard, and deal more honestly with methods that can't support that requirement. Make a full code-review pass over the tablesample additions, and fix assorted bugs, omissions, infelicities, and cosmetic issues (such as failure to put the added code stanzas in a consistent ordering). Improve EXPLAIN's output of tablesample plans, too. Back-patch to 9.5 so that we don't have to support the original API in production.
2015-07-25 14:39:00 -04:00
if (walker(rte->tablesample, context))
return true;
TABLESAMPLE, SQL Standard and extensible Add a TABLESAMPLE clause to SELECT statements that allows user to specify random BERNOULLI sampling or block level SYSTEM sampling. Implementation allows for extensible sampling functions to be written, using a standard API. Basic version follows SQLStandard exactly. Usable concrete use cases for the sampling API follow in later commits. Petr Jelinek Reviewed by Michael Paquier and Simon Riggs
2015-05-15 14:37:10 -04:00
break;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
case RTE_CTE:
Add infrastructure to support EphemeralNamedRelation references. A QueryEnvironment concept is added, which allows new types of objects to be passed into queries from parsing on through execution. At this point, the only thing implemented is a collection of EphemeralNamedRelation objects -- relations which can be referenced by name in queries, but do not exist in the catalogs. The only type of ENR implemented is NamedTuplestore, but provision is made to add more types fairly easily. An ENR can carry its own TupleDesc or reference a relation in the catalogs by relid. Although these features can be used without SPI, convenience functions are added to SPI so that ENRs can easily be used by code run through SPI. The initial use of all this is going to be transition tables in AFTER triggers, but that will be added to each PL as a separate commit. An incidental effect of this patch is to produce a more informative error message if an attempt is made to modify the contents of a CTE from a referencing DML statement. No tests previously covered that possibility, so one is added. Kevin Grittner and Thomas Munro Reviewed by Heikki Linnakangas, David Fetter, and Thomas Munro with valuable comments and suggestions from many others
2017-03-31 23:17:18 -05:00
case RTE_NAMEDTUPLESTORE:
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
/* nothing to do */
break;
case RTE_SUBQUERY:
if (!(flags & QTW_IGNORE_RT_SUBQUERIES))
if (walker(rte->subquery, context))
return true;
break;
case RTE_JOIN:
if (!(flags & QTW_IGNORE_JOINALIASES))
if (walker(rte->joinaliasvars, context))
return true;
break;
case RTE_FUNCTION:
Support multi-argument UNNEST(), and TABLE() syntax for multiple functions. This patch adds the ability to write TABLE( function1(), function2(), ...) as a single FROM-clause entry. The result is the concatenation of the first row from each function, followed by the second row from each function, etc; with NULLs inserted if any function produces fewer rows than others. This is believed to be a much more useful behavior than what Postgres currently does with multiple SRFs in a SELECT list. This syntax also provides a reasonable way to combine use of column definition lists with WITH ORDINALITY: put the column definition list inside TABLE(), where it's clear that it doesn't control the ordinality column as well. Also implement SQL-compliant multiple-argument UNNEST(), by turning UNNEST(a,b,c) into TABLE(unnest(a), unnest(b), unnest(c)). The SQL standard specifies TABLE() with only a single function, not multiple functions, and it seems to require an implicit UNNEST() which is not what this patch does. There may be something wrong with that reading of the spec, though, because if it's right then the spec's TABLE() is just a pointless alternative spelling of UNNEST(). After further review of that, we might choose to adopt a different syntax for what this patch does, but in any case this functionality seems clearly worthwhile. Andrew Gierth, reviewed by Zoltán Böszörményi and Heikki Linnakangas, and significantly revised by me
2013-11-21 19:37:02 -05:00
if (walker(rte->functions, context))
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
return true;
break;
Support XMLTABLE query expression XMLTABLE is defined by the SQL/XML standard as a feature that allows turning XML-formatted data into relational form, so that it can be used as a <table primary> in the FROM clause of a query. This new construct provides significant simplicity and performance benefit for XML data processing; what in a client-side custom implementation was reported to take 20 minutes can be executed in 400ms using XMLTABLE. (The same functionality was said to take 10 seconds using nested PostgreSQL XPath function calls, and 5 seconds using XMLReader under PL/Python). The implemented syntax deviates slightly from what the standard requires. First, the standard indicates that the PASSING clause is optional and that multiple XML input documents may be given to it; we make it mandatory and accept a single document only. Second, we don't currently support a default namespace to be specified. This implementation relies on a new executor node based on a hardcoded method table. (Because the grammar is fixed, there is no extensibility in the current approach; further constructs can be implemented on top of this such as JSON_TABLE, but they require changes to core code.) Author: Pavel Stehule, Álvaro Herrera Extensively reviewed by: Craig Ringer Discussion: https://postgr.es/m/CAFj8pRAgfzMD-LoSmnMGybD0WsEznLHWap8DO79+-GTRAPR4qA@mail.gmail.com
2017-03-08 12:39:37 -03:00
case RTE_TABLEFUNC:
if (walker(rte->tablefunc, context))
return true;
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case RTE_VALUES:
if (walker(rte->values_lists, context))
return true;
break;
}
Make security barrier views automatically updatable Views which are marked as security_barrier must have their quals applied before any user-defined quals are called, to prevent user-defined functions from being able to see rows which the security barrier view is intended to prevent them from seeing. Remove the restriction on security barrier views being automatically updatable by adding a new securityQuals list to the RTE structure which keeps track of the quals from security barrier views at each level, independently of the user-supplied quals. When RTEs are later discovered which have securityQuals populated, they are turned into subquery RTEs which are marked as security_barrier to prevent any user-supplied quals being pushed down (modulo LEAKPROOF quals). Dean Rasheed, reviewed by Craig Ringer, Simon Riggs, KaiGai Kohei
2014-04-12 21:04:58 -04:00
if (walker(rte->securityQuals, context))
return true;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
}
return false;
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
}
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
/*
* expression_tree_mutator() is designed to support routines that make a
* modified copy of an expression tree, with some nodes being added,
* removed, or replaced by new subtrees. The original tree is (normally)
* not changed. Each recursion level is responsible for returning a copy of
* (or appropriately modified substitute for) the subtree it is handed.
* A mutator routine should look like this:
*
* Node * my_mutator (Node *node, my_struct *context)
* {
* if (node == NULL)
* return NULL;
* // check for nodes that special work is required for, eg:
* if (IsA(node, Var))
* {
* ... create and return modified copy of Var node
* }
* else if (IsA(node, ...))
* {
* ... do special transformations of other node types
* }
* // for any node type not specially processed, do:
* return expression_tree_mutator(node, my_mutator, (void *) context);
* }
*
* The "context" argument points to a struct that holds whatever context
* information the mutator routine needs --- it can be used to return extra
* data gathered by the mutator, too. This argument is not touched by
* expression_tree_mutator, but it is passed down to recursive sub-invocations
* of my_mutator. The tree walk is started from a setup routine that
* fills in the appropriate context struct, calls my_mutator with the
* top-level node of the tree, and does any required post-processing.
*
* Each level of recursion must return an appropriately modified Node.
* If expression_tree_mutator() is called, it will make an exact copy
* of the given Node, but invoke my_mutator() to copy the sub-node(s)
* of that Node. In this way, my_mutator() has full control over the
* copying process but need not directly deal with expression trees
* that it has no interest in.
*
* Just as for expression_tree_walker, the node types handled by
* expression_tree_mutator include all those normally found in target lists
* and qualifier clauses during the planning stage.
*
* expression_tree_mutator will handle SubLink nodes by recursing normally
* into the "testexpr" subtree (which is an expression belonging to the outer
* plan). It will also call the mutator on the sub-Query node; however, when
* expression_tree_mutator itself is called on a Query node, it does nothing
* and returns the unmodified Query node. The net effect is that unless the
* mutator does something special at a Query node, sub-selects will not be
* visited or modified; the original sub-select will be linked to by the new
* SubLink node. Mutators that want to descend into sub-selects will usually
* do so by recognizing Query nodes and calling query_tree_mutator (below).
*
* expression_tree_mutator will handle a SubPlan node by recursing into the
* "testexpr" and the "args" list (which belong to the outer plan), but it
* will simply copy the link to the inner plan, since that's typically what
* expression tree mutators want. A mutator that wants to modify the subplan
* can force appropriate behavior by recognizing SubPlan expression nodes
* and doing the right thing.
*/
Node *
expression_tree_mutator(Node *node,
Node *(*mutator) (),
void *context)
{
/*
* The mutator has already decided not to modify the current node, but we
* must call the mutator for any sub-nodes.
*/
#define FLATCOPY(newnode, node, nodetype) \
( (newnode) = (nodetype *) palloc(sizeof(nodetype)), \
memcpy((newnode), (node), sizeof(nodetype)) )
#define CHECKFLATCOPY(newnode, node, nodetype) \
( AssertMacro(IsA((node), nodetype)), \
(newnode) = (nodetype *) palloc(sizeof(nodetype)), \
memcpy((newnode), (node), sizeof(nodetype)) )
#define MUTATE(newfield, oldfield, fieldtype) \
( (newfield) = (fieldtype) mutator((Node *) (oldfield), context) )
if (node == NULL)
return NULL;
/* Guard against stack overflow due to overly complex expressions */
check_stack_depth();
switch (nodeTag(node))
{
/*
* Primitive node types with no expression subnodes. Var and
* Const are frequent enough to deserve special cases, the others
* we just use copyObject for.
*/
case T_Var:
{
Var *var = (Var *) node;
Var *newnode;
FLATCOPY(newnode, var, Var);
return (Node *) newnode;
}
break;
case T_Const:
{
Const *oldnode = (Const *) node;
Const *newnode;
FLATCOPY(newnode, oldnode, Const);
/* XXX we don't bother with datumCopy; should we? */
return (Node *) newnode;
}
break;
case T_Param:
case T_CaseTestExpr:
Code review for NextValueExpr expression node type. Add missing infrastructure for this node type, notably in ruleutils.c where its lack could demonstrably cause EXPLAIN to fail. Add outfuncs/readfuncs support. (outfuncs support is useful today for debugging purposes. The readfuncs support may never be needed, since at present it would only matter for parallel query and NextValueExpr should never appear in a parallelizable query; but it seems like a bad idea to have a primnode type that isn't fully supported here.) Teach planner infrastructure that NextValueExpr is a volatile, parallel-unsafe, non-leaky expression node with cost cpu_operator_cost. Given its limited scope of usage, there *might* be no live bug today from the lack of that knowledge, but it's certainly going to bite us on the rear someday. Teach pg_stat_statements about the new node type, too. While at it, also teach cost_qual_eval() that MinMaxExpr, SQLValueFunction, XmlExpr, and CoerceToDomain should be charged as cpu_operator_cost. Failing to do this for SQLValueFunction was an oversight in my commit 0bb51aa96. The others are longer-standing oversights, but no time like the present to fix them. (In principle, CoerceToDomain could have cost much higher than this, but it doesn't presently seem worth trying to examine the domain's constraints here.) Modify execExprInterp.c to execute NextValueExpr as an out-of-line function; it seems quite unlikely to me that it's worth insisting that it be inlined in all expression eval methods. Besides, providing the out-of-line function doesn't stop anyone from inlining if they want to. Adjust some places where NextValueExpr support had been inserted with the aid of a dartboard rather than keeping it in the same order as elsewhere. Discussion: https://postgr.es/m/23862.1499981661@sss.pgh.pa.us
2017-07-14 15:25:43 -04:00
case T_SQLValueFunction:
case T_CoerceToDomainValue:
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_SetToDefault:
case T_CurrentOfExpr:
Identity columns This is the SQL standard-conforming variant of PostgreSQL's serial columns. It fixes a few usability issues that serial columns have: - CREATE TABLE / LIKE copies default but refers to same sequence - cannot add/drop serialness with ALTER TABLE - dropping default does not drop sequence - need to grant separate privileges to sequence - other slight weirdnesses because serial is some kind of special macro Reviewed-by: Vitaly Burovoy <vitaly.burovoy@gmail.com>
2017-04-06 08:33:16 -04:00
case T_NextValueExpr:
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_RangeTblRef:
Support ORDER BY within aggregate function calls, at long last providing a non-kluge method for controlling the order in which values are fed to an aggregate function. At the same time eliminate the old implementation restriction that DISTINCT was only supported for single-argument aggregates. Possibly release-notable behavioral change: formerly, agg(DISTINCT x) dropped null values of x unconditionally. Now, it does so only if the agg transition function is strict; otherwise nulls are treated as DISTINCT normally would, ie, you get one copy. Andrew Gierth, reviewed by Hitoshi Harada
2009-12-15 17:57:48 +00:00
case T_SortGroupClause:
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
return (Node *) copyObject(node);
WITH CHECK OPTION support for auto-updatable VIEWs For simple views which are automatically updatable, this patch allows the user to specify what level of checking should be done on records being inserted or updated. For 'LOCAL CHECK', new tuples are validated against the conditionals of the view they are being inserted into, while for 'CASCADED CHECK' the new tuples are validated against the conditionals for all views involved (from the top down). This option is part of the SQL specification. Dean Rasheed, reviewed by Pavel Stehule
2013-07-18 17:10:16 -04:00
case T_WithCheckOption:
{
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 12:12:18 -04:00
WithCheckOption *wco = (WithCheckOption *) node;
WithCheckOption *newnode;
WITH CHECK OPTION support for auto-updatable VIEWs For simple views which are automatically updatable, this patch allows the user to specify what level of checking should be done on records being inserted or updated. For 'LOCAL CHECK', new tuples are validated against the conditionals of the view they are being inserted into, while for 'CASCADED CHECK' the new tuples are validated against the conditionals for all views involved (from the top down). This option is part of the SQL specification. Dean Rasheed, reviewed by Pavel Stehule
2013-07-18 17:10:16 -04:00
FLATCOPY(newnode, wco, WithCheckOption);
MUTATE(newnode->qual, wco->qual, Node *);
return (Node *) newnode;
}
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_Aggref:
{
Aggref *aggref = (Aggref *) node;
Aggref *newnode;
FLATCOPY(newnode, aggref, Aggref);
Fix handling of argument and result datatypes for partial aggregation. When doing partial aggregation, the args list of the upper (combining) Aggref node is replaced by a Var representing the output of the partial aggregation steps, which has either the aggregate's transition data type or a serialized representation of that. However, nodeAgg.c blindly continued to use the args list as an indication of the user-level argument types. This broke resolution of polymorphic transition datatypes at executor startup (though it accidentally failed to fail for the ANYARRAY case, which is likely the only one anyone had tested). Moreover, the constructed FuncExpr passed to the finalfunc contained completely wrong information, which would have led to bogus answers or crashes for any case where the finalfunc examined that information (which is only likely to be with polymorphic aggregates using a non-polymorphic transition type). As an independent bug, apply_partialaggref_adjustment neglected to resolve a polymorphic transition datatype before assigning it as the output type of the lower-level Aggref node. This again accidentally failed to fail for ANYARRAY but would be unlikely to work in other cases. To fix the first problem, record the user-level argument types in a separate OID-list field of Aggref, and look to that rather than the args list when asking what the argument types were. (It turns out to be convenient to include any "direct" arguments in this list too, although those are not currently subject to being overwritten.) Rather than adding yet another resolve_aggregate_transtype() call to fix the second problem, add an aggtranstype field to Aggref, and store the resolved transition type OID there when the planner first computes it. (By doing this in the planner and not the parser, we can allow the aggregate's transition type to change from time to time, although no DDL support yet exists for that.) This saves nothing of consequence for simple non-polymorphic aggregates, but for polymorphic transition types we save a catalog lookup during executor startup as well as several planner lookups that are new in 9.6 due to parallel query planning. In passing, fix an error that was introduced into count_agg_clauses_walker some time ago: it was applying exprTypmod() to something that wasn't an expression node at all, but a TargetEntry. exprTypmod silently returned -1 so that there was not an obvious failure, but this broke the intended sensitivity of aggregate space consumption estimates to the typmod of varchar and similar data types. This part needs to be back-patched. Catversion bump due to change of stored Aggref nodes. Discussion: <8229.1466109074@sss.pgh.pa.us>
2016-06-17 21:44:37 -04:00
/* assume mutation doesn't change types of arguments */
newnode->aggargtypes = list_copy(aggref->aggargtypes);
Support ordered-set (WITHIN GROUP) aggregates. This patch introduces generic support for ordered-set and hypothetical-set aggregate functions, as well as implementations of the instances defined in SQL:2008 (percentile_cont(), percentile_disc(), rank(), dense_rank(), percent_rank(), cume_dist()). We also added mode() though it is not in the spec, as well as versions of percentile_cont() and percentile_disc() that can compute multiple percentile values in one pass over the data. Unlike the original submission, this patch puts full control of the sorting process in the hands of the aggregate's support functions. To allow the support functions to find out how they're supposed to sort, a new API function AggGetAggref() is added to nodeAgg.c. This allows retrieval of the aggregate call's Aggref node, which may have other uses beyond the immediate need. There is also support for ordered-set aggregates to install cleanup callback functions, so that they can be sure that infrastructure such as tuplesort objects gets cleaned up. In passing, make some fixes in the recently-added support for variadic aggregates, and make some editorial adjustments in the recent FILTER additions for aggregates. Also, simplify use of IsBinaryCoercible() by allowing it to succeed whenever the target type is ANY or ANYELEMENT. It was inconsistent that it dealt with other polymorphic target types but not these. Atri Sharma and Andrew Gierth; reviewed by Pavel Stehule and Vik Fearing, and rather heavily editorialized upon by Tom Lane
2013-12-23 16:11:35 -05:00
MUTATE(newnode->aggdirectargs, aggref->aggdirectargs, List *);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
MUTATE(newnode->args, aggref->args, List *);
Support ORDER BY within aggregate function calls, at long last providing a non-kluge method for controlling the order in which values are fed to an aggregate function. At the same time eliminate the old implementation restriction that DISTINCT was only supported for single-argument aggregates. Possibly release-notable behavioral change: formerly, agg(DISTINCT x) dropped null values of x unconditionally. Now, it does so only if the agg transition function is strict; otherwise nulls are treated as DISTINCT normally would, ie, you get one copy. Andrew Gierth, reviewed by Hitoshi Harada
2009-12-15 17:57:48 +00:00
MUTATE(newnode->aggorder, aggref->aggorder, List *);
MUTATE(newnode->aggdistinct, aggref->aggdistinct, List *);
Implement the FILTER clause for aggregate function calls. This is SQL-standard with a few extensions, namely support for subqueries and outer references in clause expressions. catversion bump due to change in Aggref and WindowFunc. David Fetter, reviewed by Dean Rasheed.
2013-07-16 20:15:36 -04:00
MUTATE(newnode->aggfilter, aggref->aggfilter, Expr *);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
return (Node *) newnode;
}
break;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingFunc:
{
pgindent run for 9.5
2015-05-23 21:35:49 -04:00
GroupingFunc *grouping = (GroupingFunc *) node;
GroupingFunc *newnode;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
FLATCOPY(newnode, grouping, GroupingFunc);
MUTATE(newnode->args, grouping->args, List *);
/*
* We assume here that mutating the arguments does not change
* the semantics, i.e. that the arguments are not mutated in a
* way that makes them semantically different from their
* previously matching expressions in the GROUP BY clause.
*
* If a mutator somehow wanted to do this, it would have to
* handle the refs and cols lists itself as appropriate.
*/
newnode->refs = list_copy(grouping->refs);
newnode->cols = list_copy(grouping->cols);
return (Node *) newnode;
}
break;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
case T_WindowFunc:
{
WindowFunc *wfunc = (WindowFunc *) node;
WindowFunc *newnode;
FLATCOPY(newnode, wfunc, WindowFunc);
MUTATE(newnode->args, wfunc->args, List *);
Implement the FILTER clause for aggregate function calls. This is SQL-standard with a few extensions, namely support for subqueries and outer references in clause expressions. catversion bump due to change in Aggref and WindowFunc. David Fetter, reviewed by Dean Rasheed.
2013-07-16 20:15:36 -04:00
MUTATE(newnode->aggfilter, wfunc->aggfilter, Expr *);
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
return (Node *) newnode;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ArrayRef:
{
ArrayRef *arrayref = (ArrayRef *) node;
ArrayRef *newnode;
FLATCOPY(newnode, arrayref, ArrayRef);
MUTATE(newnode->refupperindexpr, arrayref->refupperindexpr,
List *);
MUTATE(newnode->reflowerindexpr, arrayref->reflowerindexpr,
List *);
MUTATE(newnode->refexpr, arrayref->refexpr,
Expr *);
MUTATE(newnode->refassgnexpr, arrayref->refassgnexpr,
Expr *);
return (Node *) newnode;
}
break;
case T_FuncExpr:
{
FuncExpr *expr = (FuncExpr *) node;
FuncExpr *newnode;
FLATCOPY(newnode, expr, FuncExpr);
MUTATE(newnode->args, expr->args, List *);
return (Node *) newnode;
}
break;
Support use of function argument names to identify which actual arguments match which function parameters. The syntax uses AS, for example funcname(value AS arg1, anothervalue AS arg2) Pavel Stehule
2009-10-08 02:39:25 +00:00
case T_NamedArgExpr:
{
NamedArgExpr *nexpr = (NamedArgExpr *) node;
NamedArgExpr *newnode;
FLATCOPY(newnode, nexpr, NamedArgExpr);
MUTATE(newnode->arg, nexpr->arg, Expr *);
return (Node *) newnode;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_OpExpr:
{
OpExpr *expr = (OpExpr *) node;
OpExpr *newnode;
FLATCOPY(newnode, expr, OpExpr);
MUTATE(newnode->args, expr->args, List *);
return (Node *) newnode;
}
break;
case T_DistinctExpr:
{
DistinctExpr *expr = (DistinctExpr *) node;
DistinctExpr *newnode;
FLATCOPY(newnode, expr, DistinctExpr);
MUTATE(newnode->args, expr->args, List *);
return (Node *) newnode;
}
break;
Revise collation derivation method and expression-tree representation. All expression nodes now have an explicit output-collation field, unless they are known to only return a noncollatable data type (such as boolean or record). Also, nodes that can invoke collation-aware functions store a separate field that is the collation value to pass to the function. This avoids confusion that arises when a function has collatable inputs and noncollatable output type, or vice versa. Also, replace the parser's on-the-fly collation assignment method with a post-pass over the completed expression tree. This allows us to use a more complex (and hopefully more nearly spec-compliant) assignment rule without paying for it in extra storage in every expression node. Fix assorted bugs in the planner's handling of collations by making collation one of the defining properties of an EquivalenceClass and by converting CollateExprs into discardable RelabelType nodes during expression preprocessing.
2011-03-19 20:29:08 -04:00
case T_NullIfExpr:
{
NullIfExpr *expr = (NullIfExpr *) node;
NullIfExpr *newnode;
FLATCOPY(newnode, expr, NullIfExpr);
MUTATE(newnode->args, expr->args, List *);
return (Node *) newnode;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_ScalarArrayOpExpr:
{
ScalarArrayOpExpr *expr = (ScalarArrayOpExpr *) node;
ScalarArrayOpExpr *newnode;
FLATCOPY(newnode, expr, ScalarArrayOpExpr);
MUTATE(newnode->args, expr->args, List *);
return (Node *) newnode;
}
break;
case T_BoolExpr:
{
BoolExpr *expr = (BoolExpr *) node;
BoolExpr *newnode;
FLATCOPY(newnode, expr, BoolExpr);
MUTATE(newnode->args, expr->args, List *);
return (Node *) newnode;
}
break;
case T_SubLink:
{
SubLink *sublink = (SubLink *) node;
SubLink *newnode;
FLATCOPY(newnode, sublink, SubLink);
MUTATE(newnode->testexpr, sublink->testexpr, Node *);
/*
* Also invoke the mutator on the sublink's Query node, so it
* can recurse into the sub-query if it wants to.
*/
MUTATE(newnode->subselect, sublink->subselect, Node *);
return (Node *) newnode;
}
break;
case T_SubPlan:
{
SubPlan *subplan = (SubPlan *) node;
SubPlan *newnode;
FLATCOPY(newnode, subplan, SubPlan);
/* transform testexpr */
MUTATE(newnode->testexpr, subplan->testexpr, Node *);
/* transform args list (params to be passed to subplan) */
MUTATE(newnode->args, subplan->args, List *);
/* but not the sub-Plan itself, which is referenced as-is */
return (Node *) newnode;
}
break;
case T_AlternativeSubPlan:
{
AlternativeSubPlan *asplan = (AlternativeSubPlan *) node;
AlternativeSubPlan *newnode;
FLATCOPY(newnode, asplan, AlternativeSubPlan);
MUTATE(newnode->subplans, asplan->subplans, List *);
return (Node *) newnode;
}
break;
case T_FieldSelect:
{
FieldSelect *fselect = (FieldSelect *) node;
FieldSelect *newnode;
FLATCOPY(newnode, fselect, FieldSelect);
MUTATE(newnode->arg, fselect->arg, Expr *);
return (Node *) newnode;
}
break;
case T_FieldStore:
{
FieldStore *fstore = (FieldStore *) node;
FieldStore *newnode;
FLATCOPY(newnode, fstore, FieldStore);
MUTATE(newnode->arg, fstore->arg, Expr *);
MUTATE(newnode->newvals, fstore->newvals, List *);
newnode->fieldnums = list_copy(fstore->fieldnums);
return (Node *) newnode;
}
break;
case T_RelabelType:
{
RelabelType *relabel = (RelabelType *) node;
RelabelType *newnode;
FLATCOPY(newnode, relabel, RelabelType);
MUTATE(newnode->arg, relabel->arg, Expr *);
return (Node *) newnode;
}
break;
case T_CoerceViaIO:
{
CoerceViaIO *iocoerce = (CoerceViaIO *) node;
CoerceViaIO *newnode;
FLATCOPY(newnode, iocoerce, CoerceViaIO);
MUTATE(newnode->arg, iocoerce->arg, Expr *);
return (Node *) newnode;
}
break;
case T_ArrayCoerceExpr:
{
ArrayCoerceExpr *acoerce = (ArrayCoerceExpr *) node;
ArrayCoerceExpr *newnode;
FLATCOPY(newnode, acoerce, ArrayCoerceExpr);
MUTATE(newnode->arg, acoerce->arg, Expr *);
Support arrays over domains. Allowing arrays with a domain type as their element type was left un-done in the original domain patch, but not for any very good reason. This omission leads to such surprising results as array_agg() not working on a domain column, because the parser can't identify a suitable output type for the polymorphic aggregate. In order to fix this, first clean up the APIs of coerce_to_domain() and some internal functions in parse_coerce.c so that we consistently pass around a CoercionContext along with CoercionForm. Previously, we sometimes passed an "isExplicit" boolean flag instead, which is strictly less information; and coerce_to_domain() didn't even get that, but instead had to reverse-engineer isExplicit from CoercionForm. That's contrary to the documentation in primnodes.h that says that CoercionForm only affects display and not semantics. I don't think this change fixes any live bugs, but it makes things more consistent. The main reason for doing it though is that now build_coercion_expression() receives ccontext, which it needs in order to be able to recursively invoke coerce_to_target_type(). Next, reimplement ArrayCoerceExpr so that the node does not directly know any details of what has to be done to the individual array elements while performing the array coercion. Instead, the per-element processing is represented by a sub-expression whose input is a source array element and whose output is a target array element. This simplifies life in parse_coerce.c, because it can build that sub-expression by a recursive invocation of coerce_to_target_type(). The executor now handles the per-element processing as a compiled expression instead of hard-wired code. The main advantage of this is that we can use a single ArrayCoerceExpr to handle as many as three successive steps per element: base type conversion, typmod coercion, and domain constraint checking. The old code used two stacked ArrayCoerceExprs to handle type + typmod coercion, which was pretty inefficient, and adding yet another array deconstruction to do domain constraint checking seemed very unappetizing. In the case where we just need a single, very simple coercion function, doing this straightforwardly leads to a noticeable increase in the per-array-element runtime cost. Hence, add an additional shortcut evalfunc in execExprInterp.c that skips unnecessary overhead for that specific form of expression. The runtime speed of simple cases is within 1% or so of where it was before, while cases that previously required two levels of array processing are significantly faster. Finally, create an implicit array type for every domain type, as we do for base types, enums, etc. Everything except the array-coercion case seems to just work without further effort. Tom Lane, reviewed by Andrew Dunstan Discussion: https://postgr.es/m/9852.1499791473@sss.pgh.pa.us
2017-09-30 13:40:56 -04:00
MUTATE(newnode->elemexpr, acoerce->elemexpr, Expr *);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
return (Node *) newnode;
}
break;
case T_ConvertRowtypeExpr:
{
ConvertRowtypeExpr *convexpr = (ConvertRowtypeExpr *) node;
ConvertRowtypeExpr *newnode;
FLATCOPY(newnode, convexpr, ConvertRowtypeExpr);
MUTATE(newnode->arg, convexpr->arg, Expr *);
return (Node *) newnode;
}
break;
Split CollateClause into separate raw and analyzed node types. CollateClause is now used only in raw grammar output, and CollateExpr after parse analysis. This is for clarity and to avoid carrying collation names in post-analysis parse trees: that's both wasteful and possibly misleading, since the collation's name could be changed while the parsetree still exists. Also, clean up assorted infelicities and omissions in processing of the node type.
2011-03-11 16:27:51 -05:00
case T_CollateExpr:
{
CollateExpr *collate = (CollateExpr *) node;
CollateExpr *newnode;
FLATCOPY(newnode, collate, CollateExpr);
MUTATE(newnode->arg, collate->arg, Expr *);
return (Node *) newnode;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_CaseExpr:
{
CaseExpr *caseexpr = (CaseExpr *) node;
CaseExpr *newnode;
FLATCOPY(newnode, caseexpr, CaseExpr);
MUTATE(newnode->arg, caseexpr->arg, Expr *);
MUTATE(newnode->args, caseexpr->args, List *);
MUTATE(newnode->defresult, caseexpr->defresult, Expr *);
return (Node *) newnode;
}
break;
case T_CaseWhen:
{
CaseWhen *casewhen = (CaseWhen *) node;
CaseWhen *newnode;
FLATCOPY(newnode, casewhen, CaseWhen);
MUTATE(newnode->expr, casewhen->expr, Expr *);
MUTATE(newnode->result, casewhen->result, Expr *);
return (Node *) newnode;
}
break;
case T_ArrayExpr:
{
ArrayExpr *arrayexpr = (ArrayExpr *) node;
ArrayExpr *newnode;
FLATCOPY(newnode, arrayexpr, ArrayExpr);
MUTATE(newnode->elements, arrayexpr->elements, List *);
return (Node *) newnode;
}
break;
case T_RowExpr:
{
RowExpr *rowexpr = (RowExpr *) node;
RowExpr *newnode;
FLATCOPY(newnode, rowexpr, RowExpr);
MUTATE(newnode->args, rowexpr->args, List *);
When expanding a whole-row Var into a RowExpr during ResolveNew(), attach the column alias names of the RTE referenced by the Var to the RowExpr. This is needed to allow ruleutils.c to correctly deparse FieldSelect nodes referencing such a construct. Per my recent bug report. Adding a field to RowExpr forces initdb (because of stored rules changes) so this solution is not back-patchable; which is unfortunate because 8.2 and 8.3 have this issue. But it only affects EXPLAIN for some pretty odd corner cases, so we can probably live without a solution for the back branches.
2008-10-06 17:39:26 +00:00
/* Assume colnames needn't be duplicated */
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
return (Node *) newnode;
}
break;
case T_RowCompareExpr:
{
RowCompareExpr *rcexpr = (RowCompareExpr *) node;
RowCompareExpr *newnode;
FLATCOPY(newnode, rcexpr, RowCompareExpr);
MUTATE(newnode->largs, rcexpr->largs, List *);
MUTATE(newnode->rargs, rcexpr->rargs, List *);
return (Node *) newnode;
}
break;
case T_CoalesceExpr:
{
CoalesceExpr *coalesceexpr = (CoalesceExpr *) node;
CoalesceExpr *newnode;
FLATCOPY(newnode, coalesceexpr, CoalesceExpr);
MUTATE(newnode->args, coalesceexpr->args, List *);
return (Node *) newnode;
}
break;
case T_MinMaxExpr:
{
MinMaxExpr *minmaxexpr = (MinMaxExpr *) node;
MinMaxExpr *newnode;
FLATCOPY(newnode, minmaxexpr, MinMaxExpr);
MUTATE(newnode->args, minmaxexpr->args, List *);
return (Node *) newnode;
}
break;
case T_XmlExpr:
{
XmlExpr *xexpr = (XmlExpr *) node;
XmlExpr *newnode;
FLATCOPY(newnode, xexpr, XmlExpr);
MUTATE(newnode->named_args, xexpr->named_args, List *);
/* assume mutator does not care about arg_names */
MUTATE(newnode->args, xexpr->args, List *);
return (Node *) newnode;
}
break;
case T_NullTest:
{
NullTest *ntest = (NullTest *) node;
NullTest *newnode;
FLATCOPY(newnode, ntest, NullTest);
MUTATE(newnode->arg, ntest->arg, Expr *);
return (Node *) newnode;
}
break;
case T_BooleanTest:
{
BooleanTest *btest = (BooleanTest *) node;
BooleanTest *newnode;
FLATCOPY(newnode, btest, BooleanTest);
MUTATE(newnode->arg, btest->arg, Expr *);
return (Node *) newnode;
}
break;
case T_CoerceToDomain:
{
CoerceToDomain *ctest = (CoerceToDomain *) node;
CoerceToDomain *newnode;
FLATCOPY(newnode, ctest, CoerceToDomain);
MUTATE(newnode->arg, ctest->arg, Expr *);
return (Node *) newnode;
}
break;
case T_TargetEntry:
{
TargetEntry *targetentry = (TargetEntry *) node;
TargetEntry *newnode;
FLATCOPY(newnode, targetentry, TargetEntry);
MUTATE(newnode->expr, targetentry->expr, Expr *);
return (Node *) newnode;
}
break;
case T_Query:
/* Do nothing with a sub-Query, per discussion above */
return node;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
case T_WindowClause:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
WindowClause *wc = (WindowClause *) node;
WindowClause *newnode;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
FLATCOPY(newnode, wc, WindowClause);
MUTATE(newnode->partitionClause, wc->partitionClause, List *);
MUTATE(newnode->orderClause, wc->orderClause, List *);
Extend the set of frame options supported for window functions. This patch allows the frame to start from CURRENT ROW (in either RANGE or ROWS mode), and it also adds support for ROWS n PRECEDING and ROWS n FOLLOWING start and end points. (RANGE value PRECEDING/FOLLOWING isn't there yet --- the grammar works, but that's all.) Hitoshi Harada, reviewed by Pavel Stehule
2010-02-12 17:33:21 +00:00
MUTATE(newnode->startOffset, wc->startOffset, Node *);
MUTATE(newnode->endOffset, wc->endOffset, Node *);
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
return (Node *) newnode;
}
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_CommonTableExpr:
{
CommonTableExpr *cte = (CommonTableExpr *) node;
CommonTableExpr *newnode;
FLATCOPY(newnode, cte, CommonTableExpr);
/*
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
* Also invoke the mutator on the CTE's Query node, so it can
* recurse into the sub-query if it wants to.
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
*/
MUTATE(newnode->ctequery, cte->ctequery, Node *);
return (Node *) newnode;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_List:
{
/*
* We assume the mutator isn't interested in the list nodes
* per se, so just invoke it on each list element. NOTE: this
* would fail badly on a list with integer elements!
*/
List *resultlist;
ListCell *temp;
resultlist = NIL;
foreach(temp, (List *) node)
{
resultlist = lappend(resultlist,
mutator((Node *) lfirst(temp),
context));
}
return (Node *) resultlist;
}
break;
case T_FromExpr:
{
FromExpr *from = (FromExpr *) node;
FromExpr *newnode;
FLATCOPY(newnode, from, FromExpr);
MUTATE(newnode->fromlist, from->fromlist, List *);
MUTATE(newnode->quals, from->quals, Node *);
return (Node *) newnode;
}
break;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_OnConflictExpr:
{
pgindent run for 9.5
2015-05-23 21:35:49 -04:00
OnConflictExpr *oc = (OnConflictExpr *) node;
OnConflictExpr *newnode;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
FLATCOPY(newnode, oc, OnConflictExpr);
MUTATE(newnode->arbiterElems, oc->arbiterElems, List *);
MUTATE(newnode->arbiterWhere, oc->arbiterWhere, Node *);
MUTATE(newnode->onConflictSet, oc->onConflictSet, List *);
MUTATE(newnode->onConflictWhere, oc->onConflictWhere, Node *);
Fix ON CONFLICT bugs that manifest when used in rules. Specifically the tlist and rti of the pseudo "excluded" relation weren't properly treated by expression_tree_walker, which lead to errors when excluded was referenced inside a rule because the varnos where not properly adjusted. Similar omissions in OffsetVarNodes and expression_tree_mutator had less impact, but should obviously be fixed nonetheless. A couple tests of for ON CONFLICT UPDATE into INSERT rule bearing relations have been added. In passing I updated a couple comments.
2015-05-13 00:13:22 +02:00
MUTATE(newnode->exclRelTlist, oc->exclRelTlist, List *);
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
return (Node *) newnode;
}
break;
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
case T_MergeAction:
{
MergeAction *action = (MergeAction *) node;
MergeAction *newnode;
FLATCOPY(newnode, action, MergeAction);
MUTATE(newnode->qual, action->qual, Node *);
MUTATE(newnode->targetList, action->targetList, List *);
return (Node *) newnode;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_JoinExpr:
{
JoinExpr *join = (JoinExpr *) node;
JoinExpr *newnode;
FLATCOPY(newnode, join, JoinExpr);
MUTATE(newnode->larg, join->larg, Node *);
MUTATE(newnode->rarg, join->rarg, Node *);
MUTATE(newnode->quals, join->quals, Node *);
/* We do not mutate alias or using by default */
return (Node *) newnode;
}
break;
case T_SetOperationStmt:
{
SetOperationStmt *setop = (SetOperationStmt *) node;
SetOperationStmt *newnode;
FLATCOPY(newnode, setop, SetOperationStmt);
MUTATE(newnode->larg, setop->larg, Node *);
MUTATE(newnode->rarg, setop->rarg, Node *);
/* We do not mutate groupClauses by default */
return (Node *) newnode;
}
break;
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
case T_PlaceHolderVar:
{
PlaceHolderVar *phv = (PlaceHolderVar *) node;
PlaceHolderVar *newnode;
FLATCOPY(newnode, phv, PlaceHolderVar);
MUTATE(newnode->phexpr, phv->phexpr, Expr *);
/* Assume we need not copy the relids bitmapset */
return (Node *) newnode;
}
break;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_InferenceElem:
{
InferenceElem *inferenceelemdexpr = (InferenceElem *) node;
InferenceElem *newnode;
FLATCOPY(newnode, inferenceelemdexpr, InferenceElem);
MUTATE(newnode->expr, newnode->expr, Node *);
return (Node *) newnode;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case T_AppendRelInfo:
{
AppendRelInfo *appinfo = (AppendRelInfo *) node;
AppendRelInfo *newnode;
FLATCOPY(newnode, appinfo, AppendRelInfo);
MUTATE(newnode->translated_vars, appinfo->translated_vars, List *);
return (Node *) newnode;
}
break;
Add a concept of "placeholder" variables to the planner. These are variables that represent some expression that we desire to compute below the top level of the plan, and then let that value "bubble up" as though it were a plain Var (ie, a column value). The immediate application is to allow sub-selects to be flattened even when they are below an outer join and have non-nullable output expressions. Formerly we couldn't flatten because such an expression wouldn't properly go to NULL when evaluated above the outer join. Now, we wrap it in a PlaceHolderVar and arrange for the actual evaluation to occur below the outer join. When the resulting Var bubbles up through the join, it will be set to NULL if necessary, yielding the correct results. This fixes a planner limitation that's existed since 7.1. In future we might want to use this mechanism to re-introduce some form of Hellerstein's "expensive functions" optimization, ie place the evaluation of an expensive function at the most suitable point in the plan tree.
2008-10-21 20:42:53 +00:00
case T_PlaceHolderInfo:
{
PlaceHolderInfo *phinfo = (PlaceHolderInfo *) node;
PlaceHolderInfo *newnode;
FLATCOPY(newnode, phinfo, PlaceHolderInfo);
MUTATE(newnode->ph_var, phinfo->ph_var, PlaceHolderVar *);
/* Assume we need not copy the relids bitmapsets */
return (Node *) newnode;
}
break;
Support multi-argument UNNEST(), and TABLE() syntax for multiple functions. This patch adds the ability to write TABLE( function1(), function2(), ...) as a single FROM-clause entry. The result is the concatenation of the first row from each function, followed by the second row from each function, etc; with NULLs inserted if any function produces fewer rows than others. This is believed to be a much more useful behavior than what Postgres currently does with multiple SRFs in a SELECT list. This syntax also provides a reasonable way to combine use of column definition lists with WITH ORDINALITY: put the column definition list inside TABLE(), where it's clear that it doesn't control the ordinality column as well. Also implement SQL-compliant multiple-argument UNNEST(), by turning UNNEST(a,b,c) into TABLE(unnest(a), unnest(b), unnest(c)). The SQL standard specifies TABLE() with only a single function, not multiple functions, and it seems to require an implicit UNNEST() which is not what this patch does. There may be something wrong with that reading of the spec, though, because if it's right then the spec's TABLE() is just a pointless alternative spelling of UNNEST(). After further review of that, we might choose to adopt a different syntax for what this patch does, but in any case this functionality seems clearly worthwhile. Andrew Gierth, reviewed by Zoltán Böszörményi and Heikki Linnakangas, and significantly revised by me
2013-11-21 19:37:02 -05:00
case T_RangeTblFunction:
{
RangeTblFunction *rtfunc = (RangeTblFunction *) node;
RangeTblFunction *newnode;
FLATCOPY(newnode, rtfunc, RangeTblFunction);
MUTATE(newnode->funcexpr, rtfunc->funcexpr, Node *);
/* Assume we need not copy the coldef info lists */
return (Node *) newnode;
}
break;
Redesign tablesample method API, and do extensive code review. The original implementation of TABLESAMPLE modeled the tablesample method API on index access methods, which wasn't a good choice because, without specialized DDL commands, there's no way to build an extension that can implement a TSM. (Raw inserts into system catalogs are not an acceptable thing to do, because we can't undo them during DROP EXTENSION, nor will pg_upgrade behave sanely.) Instead adopt an API more like procedural language handlers or foreign data wrappers, wherein the only SQL-level support object needed is a single handler function identified by having a special return type. This lets us get rid of the supporting catalog altogether, so that no custom DDL support is needed for the feature. Adjust the API so that it can support non-constant tablesample arguments (the original coding assumed we could evaluate the argument expressions at ExecInitSampleScan time, which is undesirable even if it weren't outright unsafe), and discourage sampling methods from looking at invisible tuples. Make sure that the BERNOULLI and SYSTEM methods are genuinely repeatable within and across queries, as required by the SQL standard, and deal more honestly with methods that can't support that requirement. Make a full code-review pass over the tablesample additions, and fix assorted bugs, omissions, infelicities, and cosmetic issues (such as failure to put the added code stanzas in a consistent ordering). Improve EXPLAIN's output of tablesample plans, too. Back-patch to 9.5 so that we don't have to support the original API in production.
2015-07-25 14:39:00 -04:00
case T_TableSampleClause:
{
TableSampleClause *tsc = (TableSampleClause *) node;
TableSampleClause *newnode;
FLATCOPY(newnode, tsc, TableSampleClause);
MUTATE(newnode->args, tsc->args, List *);
MUTATE(newnode->repeatable, tsc->repeatable, Expr *);
return (Node *) newnode;
}
break;
Support XMLTABLE query expression XMLTABLE is defined by the SQL/XML standard as a feature that allows turning XML-formatted data into relational form, so that it can be used as a <table primary> in the FROM clause of a query. This new construct provides significant simplicity and performance benefit for XML data processing; what in a client-side custom implementation was reported to take 20 minutes can be executed in 400ms using XMLTABLE. (The same functionality was said to take 10 seconds using nested PostgreSQL XPath function calls, and 5 seconds using XMLReader under PL/Python). The implemented syntax deviates slightly from what the standard requires. First, the standard indicates that the PASSING clause is optional and that multiple XML input documents may be given to it; we make it mandatory and accept a single document only. Second, we don't currently support a default namespace to be specified. This implementation relies on a new executor node based on a hardcoded method table. (Because the grammar is fixed, there is no extensibility in the current approach; further constructs can be implemented on top of this such as JSON_TABLE, but they require changes to core code.) Author: Pavel Stehule, Álvaro Herrera Extensively reviewed by: Craig Ringer Discussion: https://postgr.es/m/CAFj8pRAgfzMD-LoSmnMGybD0WsEznLHWap8DO79+-GTRAPR4qA@mail.gmail.com
2017-03-08 12:39:37 -03:00
case T_TableFunc:
{
TableFunc *tf = (TableFunc *) node;
TableFunc *newnode;
FLATCOPY(newnode, tf, TableFunc);
MUTATE(newnode->ns_uris, tf->ns_uris, List *);
MUTATE(newnode->docexpr, tf->docexpr, Node *);
MUTATE(newnode->rowexpr, tf->rowexpr, Node *);
MUTATE(newnode->colexprs, tf->colexprs, List *);
MUTATE(newnode->coldefexprs, tf->coldefexprs, List *);
return (Node *) newnode;
}
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
default:
elog(ERROR, "unrecognized node type: %d",
(int) nodeTag(node));
break;
}
/* can't get here, but keep compiler happy */
return NULL;
}
/*
* query_tree_mutator --- initiate modification of a Query's expressions
*
* This routine exists just to reduce the number of places that need to know
* where all the expression subtrees of a Query are. Note it can be used
* for starting a walk at top level of a Query regardless of whether the
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 12:12:18 -04:00
* mutator intends to descend into subqueries. It is also useful for
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
* descending into subqueries within a mutator.
*
* Some callers want to suppress mutating of certain items in the Query,
* typically because they need to process them specially, or don't actually
* want to recurse into subqueries. This is supported by the flags argument,
* which is the bitwise OR of flag values to suppress mutating of
* indicated items. (More flag bits may be added as needed.)
*
* Normally the Query node itself is copied, but some callers want it to be
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 12:12:18 -04:00
* modified in-place; they must pass QTW_DONT_COPY_QUERY in flags. All
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
* modified substructure is safely copied in any case.
*/
Query *
query_tree_mutator(Query *query,
Node *(*mutator) (),
void *context,
int flags)
{
Assert(query != NULL && IsA(query, Query));
if (!(flags & QTW_DONT_COPY_QUERY))
{
Query *newquery;
FLATCOPY(newquery, query, Query);
query = newquery;
}
MUTATE(query->targetList, query->targetList, List *);
WITH CHECK OPTION support for auto-updatable VIEWs For simple views which are automatically updatable, this patch allows the user to specify what level of checking should be done on records being inserted or updated. For 'LOCAL CHECK', new tuples are validated against the conditionals of the view they are being inserted into, while for 'CASCADED CHECK' the new tuples are validated against the conditionals for all views involved (from the top down). This option is part of the SQL specification. Dean Rasheed, reviewed by Pavel Stehule
2013-07-18 17:10:16 -04:00
MUTATE(query->withCheckOptions, query->withCheckOptions, List *);
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
MUTATE(query->onConflict, query->onConflict, OnConflictExpr *);
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
MUTATE(query->mergeSourceTargetList, query->mergeSourceTargetList, List *);
MUTATE(query->mergeActionList, query->mergeActionList, List *);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
MUTATE(query->returningList, query->returningList, List *);
MUTATE(query->jointree, query->jointree, FromExpr *);
MUTATE(query->setOperations, query->setOperations, Node *);
MUTATE(query->havingQual, query->havingQual, Node *);
MUTATE(query->limitOffset, query->limitOffset, Node *);
MUTATE(query->limitCount, query->limitCount, Node *);
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (!(flags & QTW_IGNORE_CTE_SUBQUERIES))
MUTATE(query->cteList, query->cteList, List *);
Initial pgindent run with pg_bsd_indent version 2.0. The new indent version includes numerous fixes thanks to Piotr Stefaniak. The main changes visible in this commit are: * Nicer formatting of function-pointer declarations. * No longer unexpectedly removes spaces in expressions using casts, sizeof, or offsetof. * No longer wants to add a space in "struct structname *varname", as well as some similar cases for const- or volatile-qualified pointers. * Declarations using PG_USED_FOR_ASSERTS_ONLY are formatted more nicely. * Fixes bug where comments following declarations were sometimes placed with no space separating them from the code. * Fixes some odd decisions for comments following case labels. * Fixes some cases where comments following code were indented to less than the expected column 33. On the less good side, it now tends to put more whitespace around typedef names that are not listed in typedefs.list. This might encourage us to put more effort into typedef name collection; it's not really a bug in indent itself. There are more changes coming after this round, having to do with comment indentation and alignment of lines appearing within parentheses. I wanted to limit the size of the diffs to something that could be reviewed without one's eyes completely glazing over, so it seemed better to split up the changes as much as practical. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 14:39:04 -04:00
else /* else copy CTE list as-is */
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
query->cteList = copyObject(query->cteList);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
query->rtable = range_table_mutator(query->rtable,
mutator, context, flags);
return query;
}
/*
* range_table_mutator is just the part of query_tree_mutator that processes
* a query's rangetable. This is split out since it can be useful on
* its own.
*/
List *
range_table_mutator(List *rtable,
Node *(*mutator) (),
void *context,
int flags)
{
List *newrt = NIL;
ListCell *rt;
foreach(rt, rtable)
{
RangeTblEntry *rte = (RangeTblEntry *) lfirst(rt);
RangeTblEntry *newrte;
FLATCOPY(newrte, rte, RangeTblEntry);
switch (rte->rtekind)
{
case RTE_RELATION:
Redesign tablesample method API, and do extensive code review. The original implementation of TABLESAMPLE modeled the tablesample method API on index access methods, which wasn't a good choice because, without specialized DDL commands, there's no way to build an extension that can implement a TSM. (Raw inserts into system catalogs are not an acceptable thing to do, because we can't undo them during DROP EXTENSION, nor will pg_upgrade behave sanely.) Instead adopt an API more like procedural language handlers or foreign data wrappers, wherein the only SQL-level support object needed is a single handler function identified by having a special return type. This lets us get rid of the supporting catalog altogether, so that no custom DDL support is needed for the feature. Adjust the API so that it can support non-constant tablesample arguments (the original coding assumed we could evaluate the argument expressions at ExecInitSampleScan time, which is undesirable even if it weren't outright unsafe), and discourage sampling methods from looking at invisible tuples. Make sure that the BERNOULLI and SYSTEM methods are genuinely repeatable within and across queries, as required by the SQL standard, and deal more honestly with methods that can't support that requirement. Make a full code-review pass over the tablesample additions, and fix assorted bugs, omissions, infelicities, and cosmetic issues (such as failure to put the added code stanzas in a consistent ordering). Improve EXPLAIN's output of tablesample plans, too. Back-patch to 9.5 so that we don't have to support the original API in production.
2015-07-25 14:39:00 -04:00
MUTATE(newrte->tablesample, rte->tablesample,
TableSampleClause *);
/* we don't bother to copy eref, aliases, etc; OK? */
TABLESAMPLE, SQL Standard and extensible Add a TABLESAMPLE clause to SELECT statements that allows user to specify random BERNOULLI sampling or block level SYSTEM sampling. Implementation allows for extensible sampling functions to be written, using a standard API. Basic version follows SQLStandard exactly. Usable concrete use cases for the sampling API follow in later commits. Petr Jelinek Reviewed by Michael Paquier and Simon Riggs
2015-05-15 14:37:10 -04:00
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case RTE_CTE:
Add infrastructure to support EphemeralNamedRelation references. A QueryEnvironment concept is added, which allows new types of objects to be passed into queries from parsing on through execution. At this point, the only thing implemented is a collection of EphemeralNamedRelation objects -- relations which can be referenced by name in queries, but do not exist in the catalogs. The only type of ENR implemented is NamedTuplestore, but provision is made to add more types fairly easily. An ENR can carry its own TupleDesc or reference a relation in the catalogs by relid. Although these features can be used without SPI, convenience functions are added to SPI so that ENRs can easily be used by code run through SPI. The initial use of all this is going to be transition tables in AFTER triggers, but that will be added to each PL as a separate commit. An incidental effect of this patch is to produce a more informative error message if an attempt is made to modify the contents of a CTE from a referencing DML statement. No tests previously covered that possibility, so one is added. Kevin Grittner and Thomas Munro Reviewed by Heikki Linnakangas, David Fetter, and Thomas Munro with valuable comments and suggestions from many others
2017-03-31 23:17:18 -05:00
case RTE_NAMEDTUPLESTORE:
Redesign tablesample method API, and do extensive code review. The original implementation of TABLESAMPLE modeled the tablesample method API on index access methods, which wasn't a good choice because, without specialized DDL commands, there's no way to build an extension that can implement a TSM. (Raw inserts into system catalogs are not an acceptable thing to do, because we can't undo them during DROP EXTENSION, nor will pg_upgrade behave sanely.) Instead adopt an API more like procedural language handlers or foreign data wrappers, wherein the only SQL-level support object needed is a single handler function identified by having a special return type. This lets us get rid of the supporting catalog altogether, so that no custom DDL support is needed for the feature. Adjust the API so that it can support non-constant tablesample arguments (the original coding assumed we could evaluate the argument expressions at ExecInitSampleScan time, which is undesirable even if it weren't outright unsafe), and discourage sampling methods from looking at invisible tuples. Make sure that the BERNOULLI and SYSTEM methods are genuinely repeatable within and across queries, as required by the SQL standard, and deal more honestly with methods that can't support that requirement. Make a full code-review pass over the tablesample additions, and fix assorted bugs, omissions, infelicities, and cosmetic issues (such as failure to put the added code stanzas in a consistent ordering). Improve EXPLAIN's output of tablesample plans, too. Back-patch to 9.5 so that we don't have to support the original API in production.
2015-07-25 14:39:00 -04:00
/* nothing to do */
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
case RTE_SUBQUERY:
if (!(flags & QTW_IGNORE_RT_SUBQUERIES))
{
CHECKFLATCOPY(newrte->subquery, rte->subquery, Query);
MUTATE(newrte->subquery, newrte->subquery, Query *);
}
else
{
/* else, copy RT subqueries as-is */
newrte->subquery = copyObject(rte->subquery);
}
break;
case RTE_JOIN:
if (!(flags & QTW_IGNORE_JOINALIASES))
MUTATE(newrte->joinaliasvars, rte->joinaliasvars, List *);
else
{
/* else, copy join aliases as-is */
newrte->joinaliasvars = copyObject(rte->joinaliasvars);
}
break;
case RTE_FUNCTION:
Support multi-argument UNNEST(), and TABLE() syntax for multiple functions. This patch adds the ability to write TABLE( function1(), function2(), ...) as a single FROM-clause entry. The result is the concatenation of the first row from each function, followed by the second row from each function, etc; with NULLs inserted if any function produces fewer rows than others. This is believed to be a much more useful behavior than what Postgres currently does with multiple SRFs in a SELECT list. This syntax also provides a reasonable way to combine use of column definition lists with WITH ORDINALITY: put the column definition list inside TABLE(), where it's clear that it doesn't control the ordinality column as well. Also implement SQL-compliant multiple-argument UNNEST(), by turning UNNEST(a,b,c) into TABLE(unnest(a), unnest(b), unnest(c)). The SQL standard specifies TABLE() with only a single function, not multiple functions, and it seems to require an implicit UNNEST() which is not what this patch does. There may be something wrong with that reading of the spec, though, because if it's right then the spec's TABLE() is just a pointless alternative spelling of UNNEST(). After further review of that, we might choose to adopt a different syntax for what this patch does, but in any case this functionality seems clearly worthwhile. Andrew Gierth, reviewed by Zoltán Böszörményi and Heikki Linnakangas, and significantly revised by me
2013-11-21 19:37:02 -05:00
MUTATE(newrte->functions, rte->functions, List *);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
break;
Support XMLTABLE query expression XMLTABLE is defined by the SQL/XML standard as a feature that allows turning XML-formatted data into relational form, so that it can be used as a <table primary> in the FROM clause of a query. This new construct provides significant simplicity and performance benefit for XML data processing; what in a client-side custom implementation was reported to take 20 minutes can be executed in 400ms using XMLTABLE. (The same functionality was said to take 10 seconds using nested PostgreSQL XPath function calls, and 5 seconds using XMLReader under PL/Python). The implemented syntax deviates slightly from what the standard requires. First, the standard indicates that the PASSING clause is optional and that multiple XML input documents may be given to it; we make it mandatory and accept a single document only. Second, we don't currently support a default namespace to be specified. This implementation relies on a new executor node based on a hardcoded method table. (Because the grammar is fixed, there is no extensibility in the current approach; further constructs can be implemented on top of this such as JSON_TABLE, but they require changes to core code.) Author: Pavel Stehule, Álvaro Herrera Extensively reviewed by: Craig Ringer Discussion: https://postgr.es/m/CAFj8pRAgfzMD-LoSmnMGybD0WsEznLHWap8DO79+-GTRAPR4qA@mail.gmail.com
2017-03-08 12:39:37 -03:00
case RTE_TABLEFUNC:
MUTATE(newrte->tablefunc, rte->tablefunc, TableFunc *);
break;
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
case RTE_VALUES:
MUTATE(newrte->values_lists, rte->values_lists, List *);
break;
}
Make security barrier views automatically updatable Views which are marked as security_barrier must have their quals applied before any user-defined quals are called, to prevent user-defined functions from being able to see rows which the security barrier view is intended to prevent them from seeing. Remove the restriction on security barrier views being automatically updatable by adding a new securityQuals list to the RTE structure which keeps track of the quals from security barrier views at each level, independently of the user-supplied quals. When RTEs are later discovered which have securityQuals populated, they are turned into subquery RTEs which are marked as security_barrier to prevent any user-supplied quals being pushed down (modulo LEAKPROOF quals). Dean Rasheed, reviewed by Craig Ringer, Simon Riggs, KaiGai Kohei
2014-04-12 21:04:58 -04:00
MUTATE(newrte->securityQuals, rte->securityQuals, List *);
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
newrt = lappend(newrt, newrte);
}
return newrt;
}
/*
* query_or_expression_tree_walker --- hybrid form
*
* This routine will invoke query_tree_walker if called on a Query node,
* else will invoke the walker directly. This is a useful way of starting
* the recursion when the walker's normal change of state is not appropriate
* for the outermost Query node.
*/
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
bool
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
query_or_expression_tree_walker(Node *node,
bool (*walker) (),
void *context,
int flags)
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
{
Move exprType(), exprTypmod(), expression_tree_walker(), and related routines into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside the backend. There's probably more that should be done along this line, but this is a start anyway.
2008-08-25 22:42:34 +00:00
if (node && IsA(node, Query))
return query_tree_walker((Query *) node,
walker,
context,
flags);
else
return walker(node, context);
}
/*
* query_or_expression_tree_mutator --- hybrid form
*
* This routine will invoke query_tree_mutator if called on a Query node,
* else will invoke the mutator directly. This is a useful way of starting
* the recursion when the mutator's normal change of state is not appropriate
* for the outermost Query node.
*/
Node *
query_or_expression_tree_mutator(Node *node,
Node *(*mutator) (),
void *context,
int flags)
{
if (node && IsA(node, Query))
return (Node *) query_tree_mutator((Query *) node,
mutator,
context,
flags);
else
return mutator(node, context);
Postgres95 1.01 Distribution - Virgin Sources
1996-07-09 06:22:35 +00:00
}
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
/*
* raw_expression_tree_walker --- walk raw parse trees
*
* This has exactly the same API as expression_tree_walker, but instead of
* walking post-analysis parse trees, it knows how to walk the node types
* found in raw grammar output. (There is not currently any need for a
* combined walker, so we keep them separate in the name of efficiency.)
* Unlike expression_tree_walker, there is no special rule about query
* boundaries: we descend to everything that's possibly interesting.
*
Add support for more extensive testing of raw_expression_tree_walker(). If RAW_EXPRESSION_COVERAGE_TEST is defined, do a no-op tree walk over every basic DML statement submitted to parse analysis. If we'd had this in place earlier, bug #14153 would have been caught by buildfarm testing. The difficulty is that raw_expression_tree_walker() is only used in limited cases involving CTEs (particularly recursive ones), so it's very easy for an oversight in it to not be noticed during testing of a seemingly-unrelated feature. The type of error we can expect to catch with this is complete omission of a node type from raw_expression_tree_walker(), and perhaps also recursion into a field that doesn't contain a node tree, though that would be an unlikely mistake. It won't catch failure to add new fields that need to be recursed into, unfortunately. I'll go enable this on one or two of my own buildfarm animals once bug #14153 is dealt with. Discussion: <27861.1464040417@sss.pgh.pa.us>
2016-05-23 19:08:26 -04:00
* Currently, the node type coverage here extends only to DML statements
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
* (SELECT/INSERT/UPDATE/DELETE/MERGE) and nodes that can appear in them,
* because this is used mainly during analysis of CTEs, and only DML
* statements can appear in CTEs.
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
*/
bool
Code review for protransform patches. Fix loss of previous expression-simplification work when a transform function fires: we must not simply revert to untransformed input tree. Instead build a dummy FuncExpr node to pass to the transform function. This has the additional advantage of providing a simpler, more uniform API for transform functions. Move documentation to a somewhat less buried spot, relocate some poorly-placed code, be more wary of null constants and invalid typmod values, add an opr_sanity check on protransform function signatures, and some other minor cosmetic adjustments. Note: although this patch touches pg_proc.h, no need for catversion bump, because the changes are cosmetic and don't actually change the intended catalog contents.
2012-03-23 17:29:57 -04:00
raw_expression_tree_walker(Node *node,
bool (*walker) (),
void *context)
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
{
ListCell *temp;
/*
* The walker has already visited the current node, and so we need only
* recurse into any sub-nodes it has.
*/
if (node == NULL)
return false;
/* Guard against stack overflow due to overly complex expressions */
check_stack_depth();
switch (nodeTag(node))
{
case T_SetToDefault:
case T_CurrentOfExpr:
Improve parsetree representation of special functions such as CURRENT_DATE. We implement a dozen or so parameterless functions that the SQL standard defines special syntax for. Up to now, that was done by converting them into more or less ad-hoc constructs such as "'now'::text::date". That's messy for multiple reasons: it exposes what should be implementation details to users, and performance is worse than it needs to be in several cases. To improve matters, invent a new expression node type SQLValueFunction that can represent any of these parameterless functions. Bump catversion because this changes stored parsetrees for rules. Discussion: <30058.1463091294@sss.pgh.pa.us>
2016-08-16 20:33:01 -04:00
case T_SQLValueFunction:
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_Integer:
case T_Float:
case T_String:
case T_BitString:
case T_Null:
case T_ParamRef:
case T_A_Const:
case T_A_Star:
/* primitive node types with no subnodes */
break;
case T_Alias:
/* we assume the colnames list isn't interesting */
break;
case T_RangeVar:
return walker(((RangeVar *) node)->alias, context);
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingFunc:
return walker(((GroupingFunc *) node)->args, context);
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_SubLink:
{
SubLink *sublink = (SubLink *) node;
if (walker(sublink->testexpr, context))
return true;
/* we assume the operName is not interesting */
if (walker(sublink->subselect, context))
return true;
}
break;
case T_CaseExpr:
{
CaseExpr *caseexpr = (CaseExpr *) node;
if (walker(caseexpr->arg, context))
return true;
/* we assume walker doesn't care about CaseWhens, either */
foreach(temp, caseexpr->args)
{
Improve castNode notation by introducing list-extraction-specific variants. This extends the castNode() notation introduced by commit 5bcab1114 to provide, in one step, extraction of a list cell's pointer and coercion to a concrete node type. For example, "lfirst_node(Foo, lc)" is the same as "castNode(Foo, lfirst(lc))". Almost half of the uses of castNode that have appeared so far include a list extraction call, so this is pretty widely useful, and it saves a few more keystrokes compared to the old way. As with the previous patch, back-patch the addition of these macros to pg_list.h, so that the notation will be available when back-patching. Patch by me, after an idea of Andrew Gierth's. Discussion: https://postgr.es/m/14197.1491841216@sss.pgh.pa.us
2017-04-10 13:51:29 -04:00
CaseWhen *when = lfirst_node(CaseWhen, temp);
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(when->expr, context))
return true;
if (walker(when->result, context))
return true;
}
if (walker(caseexpr->defresult, context))
return true;
}
break;
case T_RowExpr:
When expanding a whole-row Var into a RowExpr during ResolveNew(), attach the column alias names of the RTE referenced by the Var to the RowExpr. This is needed to allow ruleutils.c to correctly deparse FieldSelect nodes referencing such a construct. Per my recent bug report. Adding a field to RowExpr forces initdb (because of stored rules changes) so this solution is not back-patchable; which is unfortunate because 8.2 and 8.3 have this issue. But it only affects EXPLAIN for some pretty odd corner cases, so we can probably live without a solution for the back branches.
2008-10-06 17:39:26 +00:00
/* Assume colnames isn't interesting */
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
return walker(((RowExpr *) node)->args, context);
case T_CoalesceExpr:
return walker(((CoalesceExpr *) node)->args, context);
case T_MinMaxExpr:
return walker(((MinMaxExpr *) node)->args, context);
case T_XmlExpr:
{
XmlExpr *xexpr = (XmlExpr *) node;
if (walker(xexpr->named_args, context))
return true;
/* we assume walker doesn't care about arg_names */
if (walker(xexpr->args, context))
return true;
}
break;
case T_NullTest:
return walker(((NullTest *) node)->arg, context);
case T_BooleanTest:
return walker(((BooleanTest *) node)->arg, context);
case T_JoinExpr:
{
JoinExpr *join = (JoinExpr *) node;
if (walker(join->larg, context))
return true;
if (walker(join->rarg, context))
return true;
if (walker(join->quals, context))
return true;
if (walker(join->alias, context))
return true;
/* using list is deemed uninteresting */
}
break;
case T_IntoClause:
{
IntoClause *into = (IntoClause *) node;
if (walker(into->rel, context))
return true;
/* colNames, options are deemed uninteresting */
Clean up the mess around EXPLAIN and materialized views. Revert the matview-related changes in explain.c's API, as per recent complaint from Robert Haas. The reason for these appears to have been principally some ill-considered choices around having intorel_startup do what ought to be parse-time checking, plus a poor arrangement for passing it the view parsetree it needs to store into pg_rewrite when creating a materialized view. Do the latter by having parse analysis stick a copy into the IntoClause, instead of doing it at runtime. (On the whole, I seriously question the choice to represent CREATE MATERIALIZED VIEW as a variant of SELECT INTO/CREATE TABLE AS, because that means injecting even more complexity into what was already a horrid legacy kluge. However, I didn't go so far as to rethink that choice ... yet.) I also moved several error checks into matview parse analysis, and made the check for external Params in a matview more accurate. In passing, clean things up a bit more around interpretOidsOption(), and fix things so that we can use that to force no-oids for views, sequences, etc, thereby eliminating the need to cons up "oids = false" options when creating them. catversion bump due to change in IntoClause. (I wonder though if we really need readfuncs/outfuncs support for IntoClause anymore.)
2013-04-12 19:25:20 -04:00
/* viewQuery should be null in raw parsetree, but check it */
if (walker(into->viewQuery, context))
return true;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
}
break;
case T_List:
foreach(temp, (List *) node)
{
if (walker((Node *) lfirst(temp), context))
return true;
}
break;
Support data-modifying commands (INSERT/UPDATE/DELETE) in WITH. This patch implements data-modifying WITH queries according to the semantics that the updates all happen with the same command counter value, and in an unspecified order. Therefore one WITH clause can't see the effects of another, nor can the outer query see the effects other than through the RETURNING values. And attempts to do conflicting updates will have unpredictable results. We'll need to document all that. This commit just fixes the code; documentation updates are waiting on author. Marko Tiikkaja and Hitoshi Harada
2011-02-25 18:56:23 -05:00
case T_InsertStmt:
{
InsertStmt *stmt = (InsertStmt *) node;
if (walker(stmt->relation, context))
return true;
if (walker(stmt->cols, context))
return true;
if (walker(stmt->selectStmt, context))
return true;
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
if (walker(stmt->onConflictClause, context))
return true;
Support data-modifying commands (INSERT/UPDATE/DELETE) in WITH. This patch implements data-modifying WITH queries according to the semantics that the updates all happen with the same command counter value, and in an unspecified order. Therefore one WITH clause can't see the effects of another, nor can the outer query see the effects other than through the RETURNING values. And attempts to do conflicting updates will have unpredictable results. We'll need to document all that. This commit just fixes the code; documentation updates are waiting on author. Marko Tiikkaja and Hitoshi Harada
2011-02-25 18:56:23 -05:00
if (walker(stmt->returningList, context))
return true;
if (walker(stmt->withClause, context))
return true;
}
break;
case T_DeleteStmt:
{
DeleteStmt *stmt = (DeleteStmt *) node;
if (walker(stmt->relation, context))
return true;
if (walker(stmt->usingClause, context))
return true;
if (walker(stmt->whereClause, context))
return true;
if (walker(stmt->returningList, context))
return true;
if (walker(stmt->withClause, context))
return true;
}
break;
case T_UpdateStmt:
{
UpdateStmt *stmt = (UpdateStmt *) node;
if (walker(stmt->relation, context))
return true;
if (walker(stmt->targetList, context))
return true;
if (walker(stmt->whereClause, context))
return true;
if (walker(stmt->fromClause, context))
return true;
if (walker(stmt->returningList, context))
return true;
if (walker(stmt->withClause, context))
return true;
}
break;
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
case T_MergeStmt:
{
MergeStmt *stmt = (MergeStmt *) node;
if (walker(stmt->relation, context))
return true;
if (walker(stmt->source_relation, context))
return true;
if (walker(stmt->join_condition, context))
return true;
Improve parse representation for MERGE Separation of parser data structures from executor, as requested by Tom Lane. Further improvements possible. While there, implement error for multiple VALUES clauses via parser to allow line number of error, as requested by Andres Freund. Author: Pavan Deolasee Discussion: https://www.postgresql.org/message-id/CABOikdPpqjectFchg0FyTOpsGXyPoqwgC==OLKWuxgBOsrDDZw@mail.gmail.com
2018-04-06 09:38:59 +01:00
if (walker(stmt->mergeWhenClauses, context))
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
return true;
WITH support in MERGE Author: Peter Geoghegan Recursive support removed, no tests Docs added by me
2018-04-03 12:13:59 +01:00
if (walker(stmt->withClause, context))
return true;
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
}
break;
Improve parse representation for MERGE Separation of parser data structures from executor, as requested by Tom Lane. Further improvements possible. While there, implement error for multiple VALUES clauses via parser to allow line number of error, as requested by Andres Freund. Author: Pavan Deolasee Discussion: https://www.postgresql.org/message-id/CABOikdPpqjectFchg0FyTOpsGXyPoqwgC==OLKWuxgBOsrDDZw@mail.gmail.com
2018-04-06 09:38:59 +01:00
case T_MergeWhenClause:
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
{
Improve parse representation for MERGE Separation of parser data structures from executor, as requested by Tom Lane. Further improvements possible. While there, implement error for multiple VALUES clauses via parser to allow line number of error, as requested by Andres Freund. Author: Pavan Deolasee Discussion: https://www.postgresql.org/message-id/CABOikdPpqjectFchg0FyTOpsGXyPoqwgC==OLKWuxgBOsrDDZw@mail.gmail.com
2018-04-06 09:38:59 +01:00
MergeWhenClause *mergeWhenClause = (MergeWhenClause *) node;
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
Improve parse representation for MERGE Separation of parser data structures from executor, as requested by Tom Lane. Further improvements possible. While there, implement error for multiple VALUES clauses via parser to allow line number of error, as requested by Andres Freund. Author: Pavan Deolasee Discussion: https://www.postgresql.org/message-id/CABOikdPpqjectFchg0FyTOpsGXyPoqwgC==OLKWuxgBOsrDDZw@mail.gmail.com
2018-04-06 09:38:59 +01:00
if (walker(mergeWhenClause->condition, context))
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
return true;
Improve parse representation for MERGE Separation of parser data structures from executor, as requested by Tom Lane. Further improvements possible. While there, implement error for multiple VALUES clauses via parser to allow line number of error, as requested by Andres Freund. Author: Pavan Deolasee Discussion: https://www.postgresql.org/message-id/CABOikdPpqjectFchg0FyTOpsGXyPoqwgC==OLKWuxgBOsrDDZw@mail.gmail.com
2018-04-06 09:38:59 +01:00
if (walker(mergeWhenClause->targetList, context))
return true;
if (walker(mergeWhenClause->cols, context))
return true;
if (walker(mergeWhenClause->values, context))
MERGE SQL Command following SQL:2016 MERGE performs actions that modify rows in the target table using a source table or query. MERGE provides a single SQL statement that can conditionally INSERT/UPDATE/DELETE rows a task that would other require multiple PL statements. e.g. MERGE INTO target AS t USING source AS s ON t.tid = s.sid WHEN MATCHED AND t.balance > s.delta THEN UPDATE SET balance = t.balance - s.delta WHEN MATCHED THEN DELETE WHEN NOT MATCHED AND s.delta > 0 THEN INSERT VALUES (s.sid, s.delta) WHEN NOT MATCHED THEN DO NOTHING; MERGE works with regular and partitioned tables, including column and row security enforcement, as well as support for row, statement and transition triggers. MERGE is optimized for OLTP and is parameterizable, though also useful for large scale ETL/ELT. MERGE is not intended to be used in preference to existing single SQL commands for INSERT, UPDATE or DELETE since there is some overhead. MERGE can be used statically from PL/pgSQL. MERGE does not yet support inheritance, write rules, RETURNING clauses, updatable views or foreign tables. MERGE follows SQL Standard per the most recent SQL:2016. Includes full tests and documentation, including full isolation tests to demonstrate the concurrent behavior. This version written from scratch in 2017 by Simon Riggs, using docs and tests originally written in 2009. Later work from Pavan Deolasee has been both complex and deep, leaving the lead author credit now in his hands. Extensive discussion of concurrency from Peter Geoghegan, with thanks for the time and effort contributed. Various issues reported via sqlsmith by Andreas Seltenreich Authors: Pavan Deolasee, Simon Riggs Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs Discussion: https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com
2018-04-03 09:28:16 +01:00
return true;
}
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_SelectStmt:
{
SelectStmt *stmt = (SelectStmt *) node;
if (walker(stmt->distinctClause, context))
return true;
if (walker(stmt->intoClause, context))
return true;
if (walker(stmt->targetList, context))
return true;
if (walker(stmt->fromClause, context))
return true;
if (walker(stmt->whereClause, context))
return true;
if (walker(stmt->groupClause, context))
return true;
if (walker(stmt->havingClause, context))
return true;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
if (walker(stmt->windowClause, context))
return true;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(stmt->valuesLists, context))
return true;
if (walker(stmt->sortClause, context))
return true;
if (walker(stmt->limitOffset, context))
return true;
if (walker(stmt->limitCount, context))
return true;
if (walker(stmt->lockingClause, context))
return true;
Fix WITH attached to a nested set operation (UNION/INTERSECT/EXCEPT). Parse analysis neglected to cover the case of a WITH clause attached to an intermediate-level set operation; it only handled WITH at the top level or WITH attached to a leaf-level SELECT. Per report from Adam Mackler. In HEAD, I rearranged the order of SelectStmt's fields to put withClause with the other fields that can appear on non-leaf SelectStmts. In back branches, leave it alone to avoid a possible ABI break for third-party code. Back-patch to 8.4 where WITH support was added.
2012-07-31 17:56:21 -04:00
if (walker(stmt->withClause, context))
return true;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(stmt->larg, context))
return true;
if (walker(stmt->rarg, context))
return true;
}
break;
case T_A_Expr:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
A_Expr *expr = (A_Expr *) node;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(expr->lexpr, context))
return true;
if (walker(expr->rexpr, context))
return true;
/* operator name is deemed uninteresting */
}
break;
Avoid recursion when processing simple lists of AND'ed or OR'ed clauses. Since most of the system thinks AND and OR are N-argument expressions anyway, let's have the grammar generate a representation of that form when dealing with input like "x AND y AND z AND ...", rather than generating a deeply-nested binary tree that just has to be flattened later by the planner. This avoids stack overflow in parse analysis when dealing with queries having more than a few thousand such clauses; and in any case it removes some rather unsightly inconsistencies, since some parts of parse analysis were generating N-argument ANDs/ORs already. It's still possible to get a stack overflow with weirdly parenthesized input, such as "x AND (y AND (z AND ( ... )))", but such cases are not mainstream usage. The maximum depth of parenthesization is already limited by Bison's stack in such cases, anyway, so that the limit is probably fairly platform-independent. Patch originally by Gurjeet Singh, heavily revised by me
2014-06-16 15:55:05 -04:00
case T_BoolExpr:
{
BoolExpr *expr = (BoolExpr *) node;
if (walker(expr->args, context))
return true;
}
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_ColumnRef:
/* we assume the fields contain nothing interesting */
break;
case T_FuncCall:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
FuncCall *fcall = (FuncCall *) node;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(fcall->args, context))
return true;
Support ORDER BY within aggregate function calls, at long last providing a non-kluge method for controlling the order in which values are fed to an aggregate function. At the same time eliminate the old implementation restriction that DISTINCT was only supported for single-argument aggregates. Possibly release-notable behavioral change: formerly, agg(DISTINCT x) dropped null values of x unconditionally. Now, it does so only if the agg transition function is strict; otherwise nulls are treated as DISTINCT normally would, ie, you get one copy. Andrew Gierth, reviewed by Hitoshi Harada
2009-12-15 17:57:48 +00:00
if (walker(fcall->agg_order, context))
return true;
Implement the FILTER clause for aggregate function calls. This is SQL-standard with a few extensions, namely support for subqueries and outer references in clause expressions. catversion bump due to change in Aggref and WindowFunc. David Fetter, reviewed by Dean Rasheed.
2013-07-16 20:15:36 -04:00
if (walker(fcall->agg_filter, context))
return true;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
if (walker(fcall->over, context))
return true;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
/* function name is deemed uninteresting */
}
break;
Support use of function argument names to identify which actual arguments match which function parameters. The syntax uses AS, for example funcname(value AS arg1, anothervalue AS arg2) Pavel Stehule
2009-10-08 02:39:25 +00:00
case T_NamedArgExpr:
return walker(((NamedArgExpr *) node)->arg, context);
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_A_Indices:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
A_Indices *indices = (A_Indices *) node;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(indices->lidx, context))
return true;
if (walker(indices->uidx, context))
return true;
}
break;
case T_A_Indirection:
{
A_Indirection *indir = (A_Indirection *) node;
if (walker(indir->arg, context))
return true;
if (walker(indir->indirection, context))
return true;
}
break;
case T_A_ArrayExpr:
return walker(((A_ArrayExpr *) node)->elements, context);
case T_ResTarget:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
ResTarget *rt = (ResTarget *) node;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(rt->indirection, context))
return true;
if (walker(rt->val, context))
return true;
}
break;
Implement UPDATE tab SET (col1,col2,...) = (SELECT ...), ... This SQL-standard feature allows a sub-SELECT yielding multiple columns (but only one row) to be used to compute the new values of several columns to be updated. While the same results can be had with an independent sub-SELECT per column, such a workaround can require a great deal of duplicated computation. The standard actually says that the source for a multi-column assignment could be any row-valued expression. The implementation used here is tightly tied to our existing sub-SELECT support and can't handle other cases; the Bison grammar would have some issues with them too. However, I don't feel too bad about this since other cases can be converted into sub-SELECTs. For instance, "SET (a,b,c) = row_valued_function(x)" could be written "SET (a,b,c) = (SELECT * FROM row_valued_function(x))".
2014-06-18 13:22:25 -04:00
case T_MultiAssignRef:
return walker(((MultiAssignRef *) node)->source, context);
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_TypeCast:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
TypeCast *tc = (TypeCast *) node;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(tc->arg, context))
return true;
Make backend header files C++ safe This alters various incidental uses of C++ key words to use other similar identifiers, so that a C++ compiler won't choke outright. You still (probably) need extern "C" { }; around the inclusion of backend headers. based on a patch by Kurt Harriman <harriman@acm.org> Also add a script cpluspluscheck to check for C++ compatibility in the future. As of right now, this passes without error for me.
2009-07-16 06:33:46 +00:00
if (walker(tc->typeName, context))
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
return true;
}
break;
Per-column collation support This adds collation support for columns and domains, a COLLATE clause to override it per expression, and B-tree index support. Peter Eisentraut reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
2011-02-08 23:04:18 +02:00
case T_CollateClause:
return walker(((CollateClause *) node)->arg, context);
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_SortBy:
return walker(((SortBy *) node)->node, context);
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
case T_WindowDef:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
WindowDef *wd = (WindowDef *) node;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
if (walker(wd->partitionClause, context))
return true;
if (walker(wd->orderClause, context))
return true;
Extend the set of frame options supported for window functions. This patch allows the frame to start from CURRENT ROW (in either RANGE or ROWS mode), and it also adds support for ROWS n PRECEDING and ROWS n FOLLOWING start and end points. (RANGE value PRECEDING/FOLLOWING isn't there yet --- the grammar works, but that's all.) Hitoshi Harada, reviewed by Pavel Stehule
2010-02-12 17:33:21 +00:00
if (walker(wd->startOffset, context))
return true;
if (walker(wd->endOffset, context))
return true;
Support window functions a la SQL:2008. Hitoshi Harada, with some kibitzing from Heikki and Tom.
2008-12-28 18:54:01 +00:00
}
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_RangeSubselect:
{
RangeSubselect *rs = (RangeSubselect *) node;
if (walker(rs->subquery, context))
return true;
if (walker(rs->alias, context))
return true;
}
break;
case T_RangeFunction:
{
RangeFunction *rf = (RangeFunction *) node;
Support multi-argument UNNEST(), and TABLE() syntax for multiple functions. This patch adds the ability to write TABLE( function1(), function2(), ...) as a single FROM-clause entry. The result is the concatenation of the first row from each function, followed by the second row from each function, etc; with NULLs inserted if any function produces fewer rows than others. This is believed to be a much more useful behavior than what Postgres currently does with multiple SRFs in a SELECT list. This syntax also provides a reasonable way to combine use of column definition lists with WITH ORDINALITY: put the column definition list inside TABLE(), where it's clear that it doesn't control the ordinality column as well. Also implement SQL-compliant multiple-argument UNNEST(), by turning UNNEST(a,b,c) into TABLE(unnest(a), unnest(b), unnest(c)). The SQL standard specifies TABLE() with only a single function, not multiple functions, and it seems to require an implicit UNNEST() which is not what this patch does. There may be something wrong with that reading of the spec, though, because if it's right then the spec's TABLE() is just a pointless alternative spelling of UNNEST(). After further review of that, we might choose to adopt a different syntax for what this patch does, but in any case this functionality seems clearly worthwhile. Andrew Gierth, reviewed by Zoltán Böszörményi and Heikki Linnakangas, and significantly revised by me
2013-11-21 19:37:02 -05:00
if (walker(rf->functions, context))
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
return true;
if (walker(rf->alias, context))
return true;
Support multi-argument UNNEST(), and TABLE() syntax for multiple functions. This patch adds the ability to write TABLE( function1(), function2(), ...) as a single FROM-clause entry. The result is the concatenation of the first row from each function, followed by the second row from each function, etc; with NULLs inserted if any function produces fewer rows than others. This is believed to be a much more useful behavior than what Postgres currently does with multiple SRFs in a SELECT list. This syntax also provides a reasonable way to combine use of column definition lists with WITH ORDINALITY: put the column definition list inside TABLE(), where it's clear that it doesn't control the ordinality column as well. Also implement SQL-compliant multiple-argument UNNEST(), by turning UNNEST(a,b,c) into TABLE(unnest(a), unnest(b), unnest(c)). The SQL standard specifies TABLE() with only a single function, not multiple functions, and it seems to require an implicit UNNEST() which is not what this patch does. There may be something wrong with that reading of the spec, though, because if it's right then the spec's TABLE() is just a pointless alternative spelling of UNNEST(). After further review of that, we might choose to adopt a different syntax for what this patch does, but in any case this functionality seems clearly worthwhile. Andrew Gierth, reviewed by Zoltán Böszörményi and Heikki Linnakangas, and significantly revised by me
2013-11-21 19:37:02 -05:00
if (walker(rf->coldeflist, context))
return true;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
}
break;
Redesign tablesample method API, and do extensive code review. The original implementation of TABLESAMPLE modeled the tablesample method API on index access methods, which wasn't a good choice because, without specialized DDL commands, there's no way to build an extension that can implement a TSM. (Raw inserts into system catalogs are not an acceptable thing to do, because we can't undo them during DROP EXTENSION, nor will pg_upgrade behave sanely.) Instead adopt an API more like procedural language handlers or foreign data wrappers, wherein the only SQL-level support object needed is a single handler function identified by having a special return type. This lets us get rid of the supporting catalog altogether, so that no custom DDL support is needed for the feature. Adjust the API so that it can support non-constant tablesample arguments (the original coding assumed we could evaluate the argument expressions at ExecInitSampleScan time, which is undesirable even if it weren't outright unsafe), and discourage sampling methods from looking at invisible tuples. Make sure that the BERNOULLI and SYSTEM methods are genuinely repeatable within and across queries, as required by the SQL standard, and deal more honestly with methods that can't support that requirement. Make a full code-review pass over the tablesample additions, and fix assorted bugs, omissions, infelicities, and cosmetic issues (such as failure to put the added code stanzas in a consistent ordering). Improve EXPLAIN's output of tablesample plans, too. Back-patch to 9.5 so that we don't have to support the original API in production.
2015-07-25 14:39:00 -04:00
case T_RangeTableSample:
{
RangeTableSample *rts = (RangeTableSample *) node;
if (walker(rts->relation, context))
return true;
/* method name is deemed uninteresting */
if (walker(rts->args, context))
return true;
if (walker(rts->repeatable, context))
return true;
}
break;
Support XMLTABLE query expression XMLTABLE is defined by the SQL/XML standard as a feature that allows turning XML-formatted data into relational form, so that it can be used as a <table primary> in the FROM clause of a query. This new construct provides significant simplicity and performance benefit for XML data processing; what in a client-side custom implementation was reported to take 20 minutes can be executed in 400ms using XMLTABLE. (The same functionality was said to take 10 seconds using nested PostgreSQL XPath function calls, and 5 seconds using XMLReader under PL/Python). The implemented syntax deviates slightly from what the standard requires. First, the standard indicates that the PASSING clause is optional and that multiple XML input documents may be given to it; we make it mandatory and accept a single document only. Second, we don't currently support a default namespace to be specified. This implementation relies on a new executor node based on a hardcoded method table. (Because the grammar is fixed, there is no extensibility in the current approach; further constructs can be implemented on top of this such as JSON_TABLE, but they require changes to core code.) Author: Pavel Stehule, Álvaro Herrera Extensively reviewed by: Craig Ringer Discussion: https://postgr.es/m/CAFj8pRAgfzMD-LoSmnMGybD0WsEznLHWap8DO79+-GTRAPR4qA@mail.gmail.com
2017-03-08 12:39:37 -03:00
case T_RangeTableFunc:
{
RangeTableFunc *rtf = (RangeTableFunc *) node;
if (walker(rtf->docexpr, context))
return true;
if (walker(rtf->rowexpr, context))
return true;
if (walker(rtf->namespaces, context))
return true;
if (walker(rtf->columns, context))
return true;
if (walker(rtf->alias, context))
return true;
}
break;
case T_RangeTableFuncCol:
{
RangeTableFuncCol *rtfc = (RangeTableFuncCol *) node;
if (walker(rtfc->colexpr, context))
return true;
if (walker(rtfc->coldefexpr, context))
return true;
}
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_TypeName:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
TypeName *tn = (TypeName *) node;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
if (walker(tn->typmods, context))
return true;
if (walker(tn->arrayBounds, context))
return true;
/* type name itself is deemed uninteresting */
}
break;
case T_ColumnDef:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 14:49:15 +00:00
ColumnDef *coldef = (ColumnDef *) node;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
Make backend header files C++ safe This alters various incidental uses of C++ key words to use other similar identifiers, so that a C++ compiler won't choke outright. You still (probably) need extern "C" { }; around the inclusion of backend headers. based on a patch by Kurt Harriman <harriman@acm.org> Also add a script cpluspluscheck to check for C++ compatibility in the future. As of right now, this passes without error for me.
2009-07-16 06:33:46 +00:00
if (walker(coldef->typeName, context))
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
return true;
if (walker(coldef->raw_default, context))
return true;
More collations cleanup, from trawling for missed collation assignments. Mostly cosmetic, though I did find that generateClonedIndexStmt failed to clone the index's collations.
2011-03-26 16:35:25 -04:00
if (walker(coldef->collClause, context))
return true;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
/* for now, constraints are ignored */
}
break;
Support IndexElem in raw_expression_tree_walker(). Needed for cases in which INSERT ... ON CONFLICT appears inside a recursive CTE item. Per bug #14153 from Thomas Alton. Patch by Peter Geoghegan, slightly adjusted by me Report: <20160521232802.22598.13537@wrigleys.postgresql.org>
2016-05-23 19:23:36 -04:00
case T_IndexElem:
{
IndexElem *indelem = (IndexElem *) node;
if (walker(indelem->expr, context))
return true;
/* collation and opclass names are deemed uninteresting */
}
break;
Support GROUPING SETS, CUBE and ROLLUP. This SQL standard functionality allows to aggregate data by different GROUP BY clauses at once. Each grouping set returns rows with columns grouped by in other sets set to NULL. This could previously be achieved by doing each grouping as a separate query, conjoined by UNION ALLs. Besides being considerably more concise, grouping sets will in many cases be faster, requiring only one scan over the underlying data. The current implementation of grouping sets only supports using sorting for input. Individual sets that share a sort order are computed in one pass. If there are sets that don't share a sort order, additional sort & aggregation steps are performed. These additional passes are sourced by the previous sort step; thus avoiding repeated scans of the source data. The code is structured in a way that adding support for purely using hash aggregation or a mix of hashing and sorting is possible. Sorting was chosen to be supported first, as it is the most generic method of implementation. Instead of, as in an earlier versions of the patch, representing the chain of sort and aggregation steps as full blown planner and executor nodes, all but the first sort are performed inside the aggregation node itself. This avoids the need to do some unusual gymnastics to handle having to return aggregated and non-aggregated tuples from underlying nodes, as well as having to shut down underlying nodes early to limit memory usage. The optimizer still builds Sort/Agg node to describe each phase, but they're not part of the plan tree, but instead additional data for the aggregation node. They're a convenient and preexisting way to describe aggregation and sorting. The first (and possibly only) sort step is still performed as a separate execution step. That retains similarity with existing group by plans, makes rescans fairly simple, avoids very deep plans (leading to slow explains) and easily allows to avoid the sorting step if the underlying data is sorted by other means. A somewhat ugly side of this patch is having to deal with a grammar ambiguity between the new CUBE keyword and the cube extension/functions named cube (and rollup). To avoid breaking existing deployments of the cube extension it has not been renamed, neither has cube been made a reserved keyword. Instead precedence hacking is used to make GROUP BY cube(..) refer to the CUBE grouping sets feature, and not the function cube(). To actually group by a function cube(), unlikely as that might be, the function name has to be quoted. Needs a catversion bump because stored rules may change. Author: Andrew Gierth and Atri Sharma, with contributions from Andres Freund Reviewed-By: Andres Freund, Noah Misch, Tom Lane, Svenne Krap, Tomas Vondra, Erik Rijkers, Marti Raudsepp, Pavel Stehule Discussion: CAOeZVidmVRe2jU6aMk_5qkxnB7dfmPROzM7Ur8JPW5j8Y5X-Lw@mail.gmail.com
2015-05-16 03:40:59 +02:00
case T_GroupingSet:
return walker(((GroupingSet *) node)->content, context);
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_LockingClause:
return walker(((LockingClause *) node)->lockedRels, context);
case T_XmlSerialize:
{
XmlSerialize *xs = (XmlSerialize *) node;
if (walker(xs->expr, context))
return true;
Make backend header files C++ safe This alters various incidental uses of C++ key words to use other similar identifiers, so that a C++ compiler won't choke outright. You still (probably) need extern "C" { }; around the inclusion of backend headers. based on a patch by Kurt Harriman <harriman@acm.org> Also add a script cpluspluscheck to check for C++ compatibility in the future. As of right now, this passes without error for me.
2009-07-16 06:33:46 +00:00
if (walker(xs->typeName, context))
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
return true;
}
break;
case T_WithClause:
return walker(((WithClause *) node)->ctes, context);
Add support for INSERT ... ON CONFLICT DO NOTHING/UPDATE. The newly added ON CONFLICT clause allows to specify an alternative to raising a unique or exclusion constraint violation error when inserting. ON CONFLICT refers to constraints that can either be specified using a inference clause (by specifying the columns of a unique constraint) or by naming a unique or exclusion constraint. DO NOTHING avoids the constraint violation, without touching the pre-existing row. DO UPDATE SET ... [WHERE ...] updates the pre-existing tuple, and has access to both the tuple proposed for insertion and the existing tuple; the optional WHERE clause can be used to prevent an update from being executed. The UPDATE SET and WHERE clauses have access to the tuple proposed for insertion using the "magic" EXCLUDED alias, and to the pre-existing tuple using the table name or its alias. This feature is often referred to as upsert. This is implemented using a new infrastructure called "speculative insertion". It is an optimistic variant of regular insertion that first does a pre-check for existing tuples and then attempts an insert. If a violating tuple was inserted concurrently, the speculatively inserted tuple is deleted and a new attempt is made. If the pre-check finds a matching tuple the alternative DO NOTHING or DO UPDATE action is taken. If the insertion succeeds without detecting a conflict, the tuple is deemed inserted. To handle the possible ambiguity between the excluded alias and a table named excluded, and for convenience with long relation names, INSERT INTO now can alias its target table. Bumps catversion as stored rules change. Author: Peter Geoghegan, with significant contributions from Heikki Linnakangas and Andres Freund. Testing infrastructure by Jeff Janes. Reviewed-By: Heikki Linnakangas, Andres Freund, Robert Haas, Simon Riggs, Dean Rasheed, Stephen Frost and many others.
2015-05-08 05:31:36 +02:00
case T_InferClause:
{
InferClause *stmt = (InferClause *) node;
if (walker(stmt->indexElems, context))
return true;
if (walker(stmt->whereClause, context))
return true;
}
break;
case T_OnConflictClause:
{
OnConflictClause *stmt = (OnConflictClause *) node;
if (walker(stmt->infer, context))
return true;
if (walker(stmt->targetList, context))
return true;
if (walker(stmt->whereClause, context))
return true;
}
break;
Implement SQL-standard WITH clauses, including WITH RECURSIVE. There are some unimplemented aspects: recursive queries must use UNION ALL (should allow UNION too), and we don't have SEARCH or CYCLE clauses. These might or might not get done for 8.4, but even without them it's a pretty useful feature. There are also a couple of small loose ends and definitional quibbles, which I'll send a memo about to pgsql-hackers shortly. But let's land the patch now so we can get on with other development. Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
2008-10-04 21:56:55 +00:00
case T_CommonTableExpr:
return walker(((CommonTableExpr *) node)->ctequery, context);
default:
elog(ERROR, "unrecognized node type: %d",
(int) nodeTag(node));
break;
}
return false;
}
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
/*
* planstate_tree_walker --- walk plan state trees
*
* The walker has already visited the current node, and so we need only
* recurse into any sub-nodes it has.
*/
bool
Clean up a few parallelism-related things that pgindent wants to mangle. In nodeFuncs.c, pgindent wants to introduce spurious indentation into the definitions of planstate_tree_walker and planstate_walk_subplans. Fix that by spreading the definition out across several lines, similar to what is already done for other walker functions in that file. In execParallel.c, in the definition of SharedExecutorInstrumentation, pgindent wants to insert more whitespace between the type name and the member name. That causes it to mangle comments later on the line. Fix by moving the comments out of line. Now that we have a bit more room, add some more details that may be useful to the next person reading this code.
2016-04-27 11:29:45 -04:00
planstate_tree_walker(PlanState *planstate,
bool (*walker) (),
void *context)
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
{
Plan *plan = planstate->plan;
Teach planstate_tree_walker about custom scans. This logic was missing from ExplainPreScanNode, from which I derived planstate_tree_walker. But it shouldn't be missing, especially not from a generic walker function, so add it. KaiGai Kohei
2015-09-22 21:42:00 -04:00
ListCell *lc;
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
/* initPlan-s */
if (planstate_walk_subplans(planstate->initPlan, walker, context))
return true;
/* lefttree */
if (outerPlanState(planstate))
{
if (walker(outerPlanState(planstate), context))
return true;
}
/* righttree */
if (innerPlanState(planstate))
{
if (walker(innerPlanState(planstate), context))
return true;
}
/* special child plans */
switch (nodeTag(plan))
{
case T_ModifyTable:
if (planstate_walk_members(((ModifyTable *) plan)->plans,
Phase 3 of pgindent updates. Don't move parenthesized lines to the left, even if that means they flow past the right margin. By default, BSD indent lines up statement continuation lines that are within parentheses so that they start just to the right of the preceding left parenthesis. However, traditionally, if that resulted in the continuation line extending to the right of the desired right margin, then indent would push it left just far enough to not overrun the margin, if it could do so without making the continuation line start to the left of the current statement indent. That makes for a weird mix of indentations unless one has been completely rigid about never violating the 80-column limit. This behavior has been pretty universally panned by Postgres developers. Hence, disable it with indent's new -lpl switch, so that parenthesized lines are always lined up with the preceding left paren. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:35:54 -04:00
((ModifyTableState *) planstate)->mt_plans,
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
walker, context))
return true;
break;
case T_Append:
if (planstate_walk_members(((Append *) plan)->appendplans,
Phase 3 of pgindent updates. Don't move parenthesized lines to the left, even if that means they flow past the right margin. By default, BSD indent lines up statement continuation lines that are within parentheses so that they start just to the right of the preceding left parenthesis. However, traditionally, if that resulted in the continuation line extending to the right of the desired right margin, then indent would push it left just far enough to not overrun the margin, if it could do so without making the continuation line start to the left of the current statement indent. That makes for a weird mix of indentations unless one has been completely rigid about never violating the 80-column limit. This behavior has been pretty universally panned by Postgres developers. Hence, disable it with indent's new -lpl switch, so that parenthesized lines are always lined up with the preceding left paren. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:35:54 -04:00
((AppendState *) planstate)->appendplans,
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
walker, context))
return true;
break;
case T_MergeAppend:
if (planstate_walk_members(((MergeAppend *) plan)->mergeplans,
Phase 3 of pgindent updates. Don't move parenthesized lines to the left, even if that means they flow past the right margin. By default, BSD indent lines up statement continuation lines that are within parentheses so that they start just to the right of the preceding left parenthesis. However, traditionally, if that resulted in the continuation line extending to the right of the desired right margin, then indent would push it left just far enough to not overrun the margin, if it could do so without making the continuation line start to the left of the current statement indent. That makes for a weird mix of indentations unless one has been completely rigid about never violating the 80-column limit. This behavior has been pretty universally panned by Postgres developers. Hence, disable it with indent's new -lpl switch, so that parenthesized lines are always lined up with the preceding left paren. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:35:54 -04:00
((MergeAppendState *) planstate)->mergeplans,
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
walker, context))
return true;
break;
case T_BitmapAnd:
if (planstate_walk_members(((BitmapAnd *) plan)->bitmapplans,
Phase 3 of pgindent updates. Don't move parenthesized lines to the left, even if that means they flow past the right margin. By default, BSD indent lines up statement continuation lines that are within parentheses so that they start just to the right of the preceding left parenthesis. However, traditionally, if that resulted in the continuation line extending to the right of the desired right margin, then indent would push it left just far enough to not overrun the margin, if it could do so without making the continuation line start to the left of the current statement indent. That makes for a weird mix of indentations unless one has been completely rigid about never violating the 80-column limit. This behavior has been pretty universally panned by Postgres developers. Hence, disable it with indent's new -lpl switch, so that parenthesized lines are always lined up with the preceding left paren. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:35:54 -04:00
((BitmapAndState *) planstate)->bitmapplans,
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
walker, context))
return true;
break;
case T_BitmapOr:
if (planstate_walk_members(((BitmapOr *) plan)->bitmapplans,
Phase 3 of pgindent updates. Don't move parenthesized lines to the left, even if that means they flow past the right margin. By default, BSD indent lines up statement continuation lines that are within parentheses so that they start just to the right of the preceding left parenthesis. However, traditionally, if that resulted in the continuation line extending to the right of the desired right margin, then indent would push it left just far enough to not overrun the margin, if it could do so without making the continuation line start to the left of the current statement indent. That makes for a weird mix of indentations unless one has been completely rigid about never violating the 80-column limit. This behavior has been pretty universally panned by Postgres developers. Hence, disable it with indent's new -lpl switch, so that parenthesized lines are always lined up with the preceding left paren. This patch is much less interesting than the first round of indent changes, but also bulkier, so I thought it best to separate the effects. Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
2017-06-21 15:35:54 -04:00
((BitmapOrState *) planstate)->bitmapplans,
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
walker, context))
return true;
break;
case T_SubqueryScan:
if (walker(((SubqueryScanState *) planstate)->subplan, context))
return true;
break;
Teach planstate_tree_walker about custom scans. This logic was missing from ExplainPreScanNode, from which I derived planstate_tree_walker. But it shouldn't be missing, especially not from a generic walker function, so add it. KaiGai Kohei
2015-09-22 21:42:00 -04:00
case T_CustomScan:
pgindent run for 9.6
2016-06-09 18:02:36 -04:00
foreach(lc, ((CustomScanState *) planstate)->custom_ps)
Teach planstate_tree_walker about custom scans. This logic was missing from ExplainPreScanNode, from which I derived planstate_tree_walker. But it shouldn't be missing, especially not from a generic walker function, so add it. KaiGai Kohei
2015-09-22 21:42:00 -04:00
{
if (walker((PlanState *) lfirst(lc), context))
return true;
}
break;
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
default:
break;
}
/* subPlan-s */
if (planstate_walk_subplans(planstate->subPlan, walker, context))
return true;
return false;
}
/*
* Walk a list of SubPlans (or initPlans, which also use SubPlan nodes).
*/
static bool
Clean up a few parallelism-related things that pgindent wants to mangle. In nodeFuncs.c, pgindent wants to introduce spurious indentation into the definitions of planstate_tree_walker and planstate_walk_subplans. Fix that by spreading the definition out across several lines, similar to what is already done for other walker functions in that file. In execParallel.c, in the definition of SharedExecutorInstrumentation, pgindent wants to insert more whitespace between the type name and the member name. That causes it to mangle comments later on the line. Fix by moving the comments out of line. Now that we have a bit more room, add some more details that may be useful to the next person reading this code.
2016-04-27 11:29:45 -04:00
planstate_walk_subplans(List *plans,
bool (*walker) (),
void *context)
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
{
ListCell *lc;
foreach(lc, plans)
{
Improve castNode notation by introducing list-extraction-specific variants. This extends the castNode() notation introduced by commit 5bcab1114 to provide, in one step, extraction of a list cell's pointer and coercion to a concrete node type. For example, "lfirst_node(Foo, lc)" is the same as "castNode(Foo, lfirst(lc))". Almost half of the uses of castNode that have appeared so far include a list extraction call, so this is pretty widely useful, and it saves a few more keystrokes compared to the old way. As with the previous patch, back-patch the addition of these macros to pg_list.h, so that the notation will be available when back-patching. Patch by me, after an idea of Andrew Gierth's. Discussion: https://postgr.es/m/14197.1491841216@sss.pgh.pa.us
2017-04-10 13:51:29 -04:00
SubPlanState *sps = lfirst_node(SubPlanState, lc);
Add new function planstate_tree_walker. ExplainPreScanNode knows how to iterate over a generic tree of plan states; factor that logic out into a separate walker function so that other code, such as upcoming patches for parallel query, can also use it. Patch by me, reviewed by Tom Lane.
2015-09-17 11:24:49 -04:00
if (walker(sps->planstate, context))
return true;
}
return false;
}
/*
* Walk the constituent plans of a ModifyTable, Append, MergeAppend,
* BitmapAnd, or BitmapOr node.
*
* Note: we don't actually need to examine the Plan list members, but
* we need the list in order to determine the length of the PlanState array.
*/
static bool
planstate_walk_members(List *plans, PlanState **planstates,
bool (*walker) (), void *context)
{
int nplans = list_length(plans);
int j;
for (j = 0; j < nplans; j++)
{
if (walker(planstates[j], context))
return true;
}
return false;
}
Reference in New Issue Copy Permalink