database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-07-27 18:02:13 +03:00
Code Activity

MDEV-26345 SELECT MIN on Spider table returns more rows than expected

Browse Source 
The Spider storage engine ignored the implicit grouping when
aggregation was converted to constant by the query optimizer.
As a result, the Spider SE returned rows more than expected.

To fix the problem, we notify the Spider SE of the existence of
the implicit grouping via Query::distinct.
This commit is contained in:
Nayuta Yanagisawa
2021-08-22 19:52:10 +00:00
parent 5a6de6f40c
commit b9730226dc
5 changed files with 102 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
sql/sql_select.cc
View File

@ -3278,13 +3278,12 @@ bool JOIN::make_aggr_tables_info()
/*
Check if the storage engine can intercept the query.
JOIN::optimize_stage2() might convert DISTINCT into GROUP BY and then
optimize away GROUP BY (group_list). In such a case, we need to notify
a storage engine supporting a group by handler of the existence of the
original DISTINCT. Thus, we set select_distinct || group_optimized_away
to Query::distinct.
The query optimizer might optimize away aggregation functins or DISTINCT.
In such a cage, we need to notify a storage engine supporting a group by
handler of the existence of the original explicit or implicit grouping.
Thus, we set select_distinct || implicit_grouping to Query::distinct.
*/
Query query= {&all_fields, select_distinct || group_optimized_away, tables_list,
Query query= {&all_fields, select_distinct || implicit_grouping, tables_list,
conds, group_list, order ? order : group_list, having};
group_by_handler *gbh= ht->create_group_by(thd, &query);