opensourcemirror
/
postgresql
mirror of git://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Releases Wiki Activity

Further tweaking of logic that decides when to materialize an uncorrelated 

subplan: do it if subplan has subplans itself, and always do it if the
subplan is an indexscan.  (I originally set it to materialize an indexscan
only if the indexqual is fairly selective, but I dunno what I was
thinking ... an unselective indexscan is still expensive ...)
This commit is contained in:
Tom Lane 2000-03-11 23:53:41 +00:00
parent bbe1ff7404
commit e8be8ffaf0
1 changed files with 14 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

37
src/backend/optimizer/plan/subselect.c
View File

@ -7,7 +7,7 @@
* Portions Copyright (c) 1994, Regents of the University of California
*
* IDENTIFICATION
* $Header: /cvsroot/pgsql/src/backend/optimizer/plan/subselect.c,v 1.29 2000/03/02 04:08:16 tgl Exp $
* $Header: /cvsroot/pgsql/src/backend/optimizer/plan/subselect.c,v 1.30 2000/03/11 23:53:41 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@ -302,8 +302,8 @@ make_subplan(SubLink *slink)
* the top of an uncorrelated/undirect correlated subplan, which lets
* us do the work of evaluating the subplan only once. We do this
* if the subplan's top plan node is anything more complicated than
* a sequential or index scan, and we do it even for those plan types
* if the qual appears selective enough to eliminate many tuples.
* a plain sequential scan, and we do it even for seqscan if the
* qual appears selective enough to eliminate many tuples.
*/
if (node->parParam == NIL)
{
@ -312,28 +312,19 @@ make_subplan(SubLink *slink)
switch (nodeTag(plan))
{
case T_SeqScan:
{
Selectivity qualsel;
if (plan->initPlan || plan->subPlan)
use_material = true;
else
{
Selectivity qualsel;
qualsel = clauselist_selectivity(subquery, plan->qual, 0);
/* Is 10% selectivity a good threshold?? */
use_material = qualsel < 0.10;
qualsel = clauselist_selectivity(subquery,
plan->qual,
0);
/* Is 10% selectivity a good threshold?? */
use_material = qualsel < 0.10;
}
break;
}
case T_IndexScan:
{
List *indxqual = ((IndexScan *) plan)->indxqualorig;
Selectivity qualsel;
qualsel = clauselist_selectivity(subquery, plan->qual, 0);
qualsel *= clauselist_selectivity(subquery, indxqual, 0);
/* Note: if index is lossy, we just double-counted the
* index selectivity. Worth fixing?
*/
/* Is 10% selectivity a good threshold?? */
use_material = qualsel < 0.10;
break;
}
case T_Material:
case T_Sort:
/* Don't add another Material node if there's one already,