database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-07-30 11:03:19 +03:00
Code Activity

Put back planner's ability to cache the results of mergejoinscansel(),

Browse Source 
which I had removed in the first cut of the EquivalenceClass rewrite to
simplify that patch a little.  But it's still important --- in a four-way
join problem mergejoinscansel() was eating about 40% of the planning time
according to gprof.  Also, improve the EquivalenceClass code to re-use
join RestrictInfos rather than generating fresh ones for each join
considered.  This saves some memory space but more importantly improves
the effectiveness of caching planning info in RestrictInfos.
This commit is contained in:
Tom Lane
2007-01-22 20:00:40 +00:00
parent 45e0736938
commit 4f06c688c7
9 changed files with 285 additions and 65 deletions
Download Patch File Download Diff File Expand all files Collapse all files

31
src/backend/utils/adt/selfuncs.c
View File

@ -15,7 +15,7 @@
*
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/utils/adt/selfuncs.c,v 1.220 2007/01/20 20:45:40 tgl Exp $
* $PostgreSQL: pgsql/src/backend/utils/adt/selfuncs.c,v 1.221 2007/01/22 20:00:40 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@ -2112,8 +2112,8 @@ icnlikejoinsel(PG_FUNCTION_ARGS)
* we can estimate how much of the input will actually be read. This
* can have a considerable impact on the cost when using indexscans.
*
* clause should be a clause already known to be mergejoinable. opfamily and
* strategy specify the sort ordering being used.
* clause should be a clause already known to be mergejoinable. opfamily,
* strategy, and nulls_first specify the sort ordering being used.
*
* *leftscan is set to the fraction of the left-hand variable expected
* to be scanned (0 to 1), and similarly *rightscan for the right-hand
@ -2121,7 +2121,7 @@ icnlikejoinsel(PG_FUNCTION_ARGS)
*/
void
mergejoinscansel(PlannerInfo *root, Node *clause,
Oid opfamily, int strategy,
Oid opfamily, int strategy, bool nulls_first,
Selectivity *leftscan,
Selectivity *rightscan)
{
@ -2214,18 +2214,39 @@ mergejoinscansel(PlannerInfo *root, Node *clause,
/*
* Now, the fraction of the left variable that will be scanned is the
* fraction that's <= the right-side maximum value. But only believe
* non-default estimates, else stick with our 1.0.
* non-default estimates, else stick with our 1.0. Also, if the sort
* order is nulls-first, we're going to have to read over any nulls too.
*/
selec = scalarineqsel(root, leop, false, &leftvar,
rightmax, op_righttype);
if (selec != DEFAULT_INEQ_SEL)
{
if (nulls_first && HeapTupleIsValid(leftvar.statsTuple))
{
Form_pg_statistic stats;
stats = (Form_pg_statistic) GETSTRUCT(leftvar.statsTuple);
selec += stats->stanullfrac;
CLAMP_PROBABILITY(selec);
}
*leftscan = selec;
}
/* And similarly for the right variable. */
selec = scalarineqsel(root, revleop, false, &rightvar,
leftmax, op_lefttype);
if (selec != DEFAULT_INEQ_SEL)
{
if (nulls_first && HeapTupleIsValid(rightvar.statsTuple))
{
Form_pg_statistic stats;
stats = (Form_pg_statistic) GETSTRUCT(rightvar.statsTuple);
selec += stats->stanullfrac;
CLAMP_PROBABILITY(selec);
}
*rightscan = selec;
}
/*
* Only one of the two fractions can really be less than 1.0; believe the