diff options
author | Tom Lane <tgl@sss.pgh.pa.us> | 1999-08-06 04:00:17 +0000 |
---|---|---|
committer | Tom Lane <tgl@sss.pgh.pa.us> | 1999-08-06 04:00:17 +0000 |
commit | e1fad50a5d362d78b9f571b71b372faaa597462a (patch) | |
tree | fb7953f8d6bb903be4d1ee6e4c3ddc3a9ea4ad26 /src/backend/optimizer/util/pathnode.c | |
parent | b7883d7e3a1d7d3d98b2bd8186ddf60011d45bdd (diff) |
Revise generation of hashjoin paths: generate one path per
hashjoinable clause, not one path for a randomly-chosen element of each
set of clauses with the same join operator. That is, if you wrote
SELECT ... WHERE t1.f1 = t2.f2 and t1.f3 = t2.f4,
and both '=' ops were the same opcode (say, all four fields are int4),
then the system would either consider hashing on f1=f2 or on f3=f4,
but it would *not* consider both possibilities. Boo hiss.
Also, revise estimation of hashjoin costs to include a penalty when the
inner join var has a high disbursion --- ie, the most common value is
pretty common. This tends to lead to badly skewed hash bucket occupancy
and way more comparisons than you'd expect on average.
I imagine that the cost calculation still needs tweaking, but at least
it generates a more reasonable plan than before on George Young's example.
Diffstat (limited to 'src/backend/optimizer/util/pathnode.c')
-rw-r--r-- | src/backend/optimizer/util/pathnode.c | 21 |
1 files changed, 11 insertions, 10 deletions
diff --git a/src/backend/optimizer/util/pathnode.c b/src/backend/optimizer/util/pathnode.c index 5a548d2a462..f1e0f5e3ae3 100644 --- a/src/backend/optimizer/util/pathnode.c +++ b/src/backend/optimizer/util/pathnode.c @@ -7,7 +7,7 @@ * * * IDENTIFICATION - * $Header: /cvsroot/pgsql/src/backend/optimizer/util/pathnode.c,v 1.52 1999/07/30 22:34:19 tgl Exp $ + * $Header: /cvsroot/pgsql/src/backend/optimizer/util/pathnode.c,v 1.53 1999/08/06 04:00:17 tgl Exp $ * *------------------------------------------------------------------------- */ @@ -550,7 +550,7 @@ create_mergejoin_path(RelOptInfo *joinrel, } /* - * create_hashjoin_path-- XXX HASH + * create_hashjoin_path * Creates a pathnode corresponding to a hash join between two relations. * * 'joinrel' is the join relation @@ -558,13 +558,14 @@ create_mergejoin_path(RelOptInfo *joinrel, * 'innersize' is the number of tuples in the inner relation * 'outerwidth' is the number of bytes per tuple in the outer relation * 'innerwidth' is the number of bytes per tuple in the inner relation - * 'outer_path' is the outer path - * 'inner_path' is the inner path - * 'pathkeys' are the new keys of the join relation + * 'outer_path' is the cheapest outer path + * 'inner_path' is the cheapest inner path + * 'pathkeys' are the path keys of the new join path * 'operator' is the hashjoin operator - * 'hashclauses' are the applicable join/restriction clauses + * 'hashclauses' is a list of the hash join clause (always a 1-element list) * 'outerkeys' are the sort varkeys for the outer relation * 'innerkeys' are the sort varkeys for the inner relation + * 'innerdisbursion' is an estimate of the disbursion of the inner hash key * */ HashPath * @@ -579,7 +580,8 @@ create_hashjoin_path(RelOptInfo *joinrel, Oid operator, List *hashclauses, List *outerkeys, - List *innerkeys) + List *innerkeys, + Cost innerdisbursion) { HashPath *pathnode = makeNode(HashPath); @@ -600,10 +602,9 @@ create_hashjoin_path(RelOptInfo *joinrel, pathnode->innerhashkeys = innerkeys; pathnode->jpath.path.path_cost = cost_hashjoin(outer_path->path_cost, inner_path->path_cost, - outerkeys, - innerkeys, outersize, innersize, - outerwidth, innerwidth); + outerwidth, innerwidth, + innerdisbursion); return pathnode; } |