>In my 3200 track library this query takes about 7.5 seconds.
>
>If anyone finds a way to make it use indexes in this case, I'm very
>interesting as it would solve some serious performance issues with some
>of the TrackStat queries.
I get an error trying to join two tables on url columns, even though both
columns are utf8_general_ci:
Illegal mix of collations (utf8_general_ci,IMPLICIT) and
(utf8_unicode_ci,IMPLICIT) for operation '='
I remember the discussion a few months back. It discussed all kinds of ideas,
such as musicbrainz id's, generating a hash from the url, etc, but in the end
didn't we determine the fastest way was to convert one column via
CONVERT(tracks.url USING UTF8), such that it could use an index?
Ultimately though, couldn't the url columns be VARCHAR charset UTF-8?
_______________________________________________
beta mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/beta
