Bug #73437 | InnoDB Fulltext Search WITH QUERY EXPANSION is very slow | ||
---|---|---|---|
Submitted: | 30 Jul 2014 15:19 | Modified: | 17 Oct 2014 6:30 |
Reporter: | Oli Sennhauser | Email Updates: | |
Status: | No Feedback | Impact on me: | |
Category: | MySQL Server: FULLTEXT search | Severity: | S3 (Non-critical) |
Version: | 5.6.19 | OS: | Linux |
Assigned to: | CPU Architecture: | Any | |
Tags: | fulltext search, innodb, performance, query expansion, slow |
[30 Jul 2014 15:19]
Oli Sennhauser
[30 Jul 2014 15:20]
Oli Sennhauser
Without QE it runs in 40 - 80 ms.
[17 Sep 2014 6:30]
MySQL Verification Team
Hello Oli, Thank you for the report and test case. I'm not seeing much difference with/without QE with the dummy data i.e 1.4s vs 6.2s(had to adjust score to >0 in place of >5). mysql> SELECT SQL_CALC_FOUND_ROWS v.ID , MATCH (v.title, v.description) AGAINST ("gezielt und getroffen" IN NATURAL LANGUAGE MODE WITH QUERY EXPANSION) AS score FROM cameras_previews v RIGHT JOIN cameras_general c ON v.VID=c.ID WHERE v.active=1 AND v.isDeleted = 0 AND v.cover != '' AND c.licenseby NOT IN (4, 8865) AND v.price > 0 AND FIND_IN_SET(314, v.excludefromdomains)=0 AND v.ID != 71378 AND (c.connector = 314 OR c.connector = 314) HAVING score > 0 ORDER BY score DESC LIMIT 0, 5; +--------+-----------------------+ | ID | score | +--------+-----------------------+ | 153215 | 0.0001641505805309862 | | 83270 | 0.0001641505805309862 | | 181748 | 0.0001641505805309862 | | 67487 | 0.0001641505805309862 | | 43340 | 0.0001641505805309862 | +--------+-----------------------+ 5 rows in set (6.21 sec) mysql> SELECT SQL_CALC_FOUND_ROWS v.ID , MATCH (v.title, v.description) AGAINST ("gezielt und getroffen" IN NATURAL LANGUAGE MODE) AS score FROM cameras_previews v RIGHT JOIN cameras_general c ON v.VID=c.ID WHERE v.active=1 AND v.isDeleted = 0 AND v.cover != '' AND c.licenseby NOT IN (4, 8865) AND v.price > 0 AND FIND_IN_SET(314, v.excludefromdomains)=0 AND v.ID != 71378 AND (c.connector = 314 OR c.connector = 314) HAVING score > 0 ORDER BY score DESC LIMIT 0, 5; +--------+------------------------+ | ID | score | +--------+------------------------+ | 153215 | 0.00004103764513274655 | | 83270 | 0.00004103764513274655 | | 181748 | 0.00004103764513274655 | | 67487 | 0.00004103764513274655 | | 43340 | 0.00004103764513274655 | +--------+------------------------+ 5 rows in set (1.41 sec) Could you please provide exact repeatable test case(may be with private note), i.e where you observed factor 100 slower? Thanks, Umesh
[18 Oct 2014 1:00]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".