mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2026-05-07 15:45:33 +02:00
Code Issues Projects Releases Packages Wiki Activity

Bug#50843: Filesort used instead of clustered index led to

Browse source 
performance degradation.

Filesort + join cache combination is preferred to full index scan because it
is usually faster. But it's not the case when the index is clustered one.

Now test_if_skip_sort_order function prefers filesort only if index isn't
clustered.

mysql-test/r/innodb_mysql.result:
  Added a test case for the bug#50843.
mysql-test/t/innodb_mysql.test:
  Added a test case for the bug#50843.
sql/sql_select.cc:
  Bug#50843: Filesort used instead of clustered index led to
  performance degradation.
  Now test_if_skip_sort_order function prefers filesort only if index isn't
  clustered.
This commit is contained in:
Evgeny Potemkin 2010-02-26 14:17:00 +03:00
commit 2d4db52eda
3 changed files with 39 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

19
sql/sql_select.cc
View file

@ -13304,12 +13304,6 @@ test_if_skip_sort_order(JOIN_TAB *tab,ORDER *order,ha_rows select_limit,
*/
if (select_limit >= table_records)
{
/*
filesort() and join cache are usually faster than reading in
index order and not using join cache
*/
if (tab->type == JT_ALL && tab->join->tables > tab->join->const_tables + 1)
DBUG_RETURN(0);
keys= *table->file->keys_to_use_for_scanning();
keys.merge(table->covering_keys);
@ -13459,6 +13453,19 @@ test_if_skip_sort_order(JOIN_TAB *tab,ORDER *order,ha_rows select_limit,
}
}
}
/*
filesort() and join cache are usually faster than reading in
index order and not using join cache, except in case that chosen
index is clustered primary key.
*/
if ((select_limit >= table_records) &&
(tab->type == JT_ALL &&
tab->join->tables > tab->join->const_tables + 1) &&
((unsigned) best_key != table->s->primary_key ||
!table->file->primary_key_is_clustered()))
DBUG_RETURN(0);
if (best_key >= 0)
{
bool quick_created= FALSE;