database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-06-27 23:21:58 +03:00
Code Activity

Rewrite of planner statistics-gathering code. ANALYZE is now available as

Browse Source 
a separate statement (though it can still be invoked as part of VACUUM, too).
pg_statistic redesigned to be more flexible about what statistics are
stored.  ANALYZE now collects a list of several of the most common values,
not just one, plus a histogram (not just the min and max values).  Random
sampling is used to make the process reasonably fast even on very large
tables.  The number of values and histogram bins collected is now
user-settable via an ALTER TABLE command.

There is more still to do; the new stats are not being used everywhere
they could be in the planner.  But the remaining changes for this project
should be localized, and the behavior is already better than before.

A not-very-related change is that sorting now makes use of btree comparison
routines if it can find one, rather than invoking '<' twice.
This commit is contained in:
Tom Lane
2001-05-07 00:43:27 +00:00
parent 9583aea9d0
commit f905d65ee3
66 changed files with 4131 additions and 2063 deletions
Download Patch File Download Diff File Expand all files Collapse all files

58
src/backend/optimizer/path/joinpath.c
View File

@ -8,15 +8,15 @@
*
*
* IDENTIFICATION
* $Header: /cvsroot/pgsql/src/backend/optimizer/path/joinpath.c,v 1.63 2001/04/15 00:48:17 tgl Exp $
* $Header: /cvsroot/pgsql/src/backend/optimizer/path/joinpath.c,v 1.64 2001/05/07 00:43:20 tgl Exp $
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
#include <sys/types.h>
#include <math.h>
#include "postgres.h"
#include "optimizer/clauses.h"
#include "optimizer/cost.h"
#include "optimizer/pathnode.h"
@ -45,7 +45,6 @@ static void hash_inner_and_outer(Query *root, RelOptInfo *joinrel,
List *restrictlist, JoinType jointype);
static Path *best_innerjoin(List *join_paths, List *outer_relid,
JoinType jointype);
static Selectivity estimate_dispersion(Query *root, Var *var);
static List *select_mergejoin_clauses(RelOptInfo *joinrel,
RelOptInfo *outerrel,
RelOptInfo *innerrel,
@ -722,7 +721,7 @@ hash_inner_and_outer(Query *root,
Expr *clause;
Var *left,
*right;
Selectivity innerdispersion;
Selectivity innerbucketsize;
List *hashclauses;
if (restrictinfo->hashjoinoperator == InvalidOid)
@ -742,34 +741,34 @@ hash_inner_and_outer(Query *root,
/*
* Check if clause is usable with these sub-rels, find inner side,
* estimate dispersion of inner var for costing purposes.
* estimate bucketsize of inner var for costing purposes.
*
* Since we tend to visit the same clauses over and over when
* planning a large query, we cache the dispersion estimates in
* planning a large query, we cache the bucketsize estimates in
* the RestrictInfo node to avoid repeated lookups of statistics.
*/
if (intMember(left->varno, outerrelids) &&
intMember(right->varno, innerrelids))
{
/* righthand side is inner */
innerdispersion = restrictinfo->right_dispersion;
if (innerdispersion < 0)
innerbucketsize = restrictinfo->right_bucketsize;
if (innerbucketsize < 0)
{
/* not cached yet */
innerdispersion = estimate_dispersion(root, right);
restrictinfo->right_dispersion = innerdispersion;
innerbucketsize = estimate_hash_bucketsize(root, right);
restrictinfo->right_bucketsize = innerbucketsize;
}
}
else if (intMember(left->varno, innerrelids) &&
intMember(right->varno, outerrelids))
{
/* lefthand side is inner */
innerdispersion = restrictinfo->left_dispersion;
if (innerdispersion < 0)
innerbucketsize = restrictinfo->left_bucketsize;
if (innerbucketsize < 0)
{
/* not cached yet */
innerdispersion = estimate_dispersion(root, left);
restrictinfo->left_dispersion = innerdispersion;
innerbucketsize = estimate_hash_bucketsize(root, left);
restrictinfo->left_bucketsize = innerbucketsize;
}
}
else
@ -790,7 +789,7 @@ hash_inner_and_outer(Query *root,
innerrel->cheapest_total_path,
restrictlist,
hashclauses,
innerdispersion));
innerbucketsize));
if (outerrel->cheapest_startup_path != outerrel->cheapest_total_path)
add_path(joinrel, (Path *)
create_hashjoin_path(joinrel,
@ -799,7 +798,7 @@ hash_inner_and_outer(Query *root,
innerrel->cheapest_total_path,
restrictlist,
hashclauses,
innerdispersion));
innerbucketsize));
}
}
@ -866,31 +865,6 @@ best_innerjoin(List *join_paths, Relids outer_relids, JoinType jointype)
return cheapest;
}
/*
* Estimate dispersion of the specified Var
*
* We use a default of 0.1 if we can't figure out anything better.
* This will typically discourage use of a hash rather strongly,
* if the inner relation is large. We do not want to hash unless
* we know that the inner rel is well-dispersed (or the alternatives
* seem much worse).
*/
static Selectivity
estimate_dispersion(Query *root, Var *var)
{
Oid relid;
if (!IsA(var, Var))
return 0.1;
relid = getrelid(var->varno, root->rtable);
if (relid == InvalidOid)
return 0.1;
return (Selectivity) get_attdispersion(relid, var->varattno, 0.1);
}
/*
* select_mergejoin_clauses
* Select mergejoin clauses that are usable for a particular join.