diff options
| author | Alexandre Alapetite <alexandre@alapetite.fr> | 2023-09-12 13:44:17 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2023-09-12 13:44:17 +0200 |
| commit | 0bf33abac8cf83dfece4ccd7f3146c373effae2c (patch) | |
| tree | cdaaa608951e9b7df0c512c52845f467f4bc2189 /CREDITS.md | |
| parent | f050a94b48499286abfb4b69f3bcb3dee5f9ea2d (diff) | |
| download | freshrss-0bf33abac8cf83dfece4ccd7f3146c373effae2c.tar.gz freshrss-0bf33abac8cf83dfece4ccd7f3146c373effae2c.zip | |
SQL: Same updateCacheUnreads for all DBs (#5648)
* SQL: Same updateCacheUnreads for all DBs
Use same SQL update request for MySQL / MariaDB than the one we already used for PostgreSQL / SQLite (i.e. using a sub-query).
Testing on a DB of 688MB with 270k entries, 199 feeds, 19 categories, using MySQL 8.1.0.
The new SQL update using a sub-query took in average 0.02s, while the old SQL update using a join took in average 0.05s. SQL cache was properly invalidated between each run. The new SQL request is thus about twice faster.
Another advantage of the SQL update using a sub-query is that it works identically in PostgreSQL, SQLite, MariaDB, MySQL, so we do need different versions anymore.
Contributes to https://github.com/FreshRSS/FreshRSS/issues/5008#issuecomment-1709755370
* Force USE INDEX
* Use same SQL methods also for markReadEntries, markReadCat
Diffstat (limited to 'CREDITS.md')
0 files changed, 0 insertions, 0 deletions