To use those indexes on title you also need to filter by log_namespace:
MariaDB [commonswiki_p]> select count(*) from logging_logindex where
log_type="thanks" and log_title="Martin_Urbanec" and log_namespace = 2;
+----------+
| count(*) |
+----------+
|        6 |
+----------+
1 row in set (0.03 sec)

On Mon, 3 Jun 2019 at 18:47, Jaime Crespo <[email protected]> wrote:

> Hello,
>
> On Mon, Jun 3, 2019 at 7:31 PM Martin Urbanec
> <[email protected]> wrote:
> >
> > Hello,
> >
> > I'm trying to run this query more fastly. Can you help me how to
> optimalize it?
> >
> > MariaDB [commonswiki_p]> select count(*) from logging_logindex where
> log_type="thanks" and log_title="Martin_Urbanec";
> > +----------+
> > | count(*) |
> > +----------+
> > |        6 |
> > +----------+
> > 1 row in set (8 min 42.87 sec)
>
> could you paste the EXPLAIN output? You can generate that with Quarry
> (or thought the command line with `SHOW EXPLAIN FOR <connection>`).
> That may help us providing alternative query methods or seeing if an
> index is missing.
>
> Without looking at the query or the table, I wonder if it is missing a
> namespace and that is why it may be doing a full scan (explain may
> help)?
>
> _______________________________________________
> Wikimedia Cloud Services mailing list
> [email protected] (formerly [email protected])
> https://lists.wikimedia.org/mailman/listinfo/cloud
_______________________________________________
Wikimedia Cloud Services mailing list
[email protected] (formerly [email protected])
https://lists.wikimedia.org/mailman/listinfo/cloud

Reply via email to