[MDEV-27453] Performance degredation when IN elements are string instead of int Created: 2022-01-09 Updated: 2022-06-11 Resolved: 2022-01-10 |
|
| Status: | Closed |
| Project: | MariaDB Server |
| Component/s: | Optimizer |
| Affects Version/s: | 10.5.12, 10.6.8 |
| Fix Version/s: | N/A |
| Type: | Bug | Priority: | Major |
| Reporter: | Panagiotis Malakoudis | Assignee: | Sergei Golubchik |
| Resolution: | Duplicate | Votes: | 0 |
| Labels: | None | ||
| Environment: |
Debian 11 |
||
| Attachments: |
|
||||||||
| Issue Links: |
|
||||||||
| Description |
|
I upgraded from Debian 10 to Debian 11. Debian 10 had mariadb 10.3.31 and Debian 11 has mariadb 10.5.12. I attach two example queries, one with quotes, one without. I will write them here cut down, with the explain output for both mariadb 10.3.31 and 10.5.12 10.3.31 with quotes 10.3.31 without quotes 10.5.12 with quotes 10.5.12 without quotes So the first problem is the different optimization for the query with the quotes, resulting in 2,3 seconds execution instead of 0,05 seconds in mariadb 10.3.31. 10.3.31 uses range while 10.5.12 does not. Second problem is that mariadb is not doing IN predicate for single quoted integers, although I thought that |
| Comments |
| Comment by Sergei Golubchik [ 2022-01-10 ] |
|
This is being fixed as |
| Comment by Panagiotis Malakoudis [ 2022-06-11 ] |
|
I still have the reported issue with 10.6.8-MariaDB-1:10.6.8+maria~bullseye mariadb.org binary distribution |