database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-07-28 23:42:10 +03:00
Code Activity

Fix cost estimates for EXISTS subqueries that are evaluated as initPlans

Browse Source 
(because they are uncorrelated with the immediate parent query).  We were
charging the full run cost to the parent node, disregarding the fact that
only one row need be fetched for EXISTS.  While this would only be a
cosmetic issue in most cases, it might possibly affect planning outcomes
if the parent query were itself a subquery to some upper query.
Per recent discussion with Steve Crawford.
This commit is contained in:
Tom Lane
2007-09-22 21:36:40 +00:00
parent 576b8903f7
commit 7125687511
3 changed files with 49 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
src/backend/optimizer/plan/subselect.c
View File

@ -7,7 +7,7 @@
* Portions Copyright (c) 1994, Regents of the University of California
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/optimizer/plan/subselect.c,v 1.124 2007/08/26 21:44:25 tgl Exp $
* $PostgreSQL: pgsql/src/backend/optimizer/plan/subselect.c,v 1.125 2007/09/22 21:36:40 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@ -18,6 +18,7 @@
#include "miscadmin.h"
#include "nodes/makefuncs.h"
#include "optimizer/clauses.h"
#include "optimizer/cost.h"
#include "optimizer/planmain.h"
#include "optimizer/planner.h"
#include "optimizer/subselect.h"
@ -266,7 +267,7 @@ make_subplan(PlannerInfo *root, SubLink *slink, Node *testexpr, bool isTopQual)
* (we're only expecting one row out, anyway).
*
* NOTE: if you change these numbers, also change cost_qual_eval_walker()
* in path/costsize.c.
* and get_initplan_cost() in path/costsize.c.
*
* XXX If an ALL/ANY subplan is uncorrelated, we may decide to hash or
* materialize its result below. In that case it would've been better to
@ -1021,7 +1022,7 @@ SS_finalize_plan(PlannerInfo *root, Plan *plan)
* have extParams that are setParams of other initPlans, so we have to
* take care of this situation explicitly.)
*
* We also add the total_cost of each initPlan to the startup cost of the
* We also add the eval cost of each initPlan to the startup cost of the
* top node. This is a conservative overestimate, since in fact each
* initPlan might be executed later than plan startup, or even not at all.
*/
@ -1041,7 +1042,7 @@ SS_finalize_plan(PlannerInfo *root, Plan *plan)
{
initSetParam = bms_add_member(initSetParam, lfirst_int(l2));
}
initplan_cost += initplan->total_cost;
initplan_cost += get_initplan_cost(root, initsubplan);
}
/* allParam must include all these params */
plan->allParam = bms_add_members(plan->allParam, initExtParam);