On Thu, Feb 23, 2012 at 5:50 PM, Daevid Vincent wrote:
> Anyone have any thoughts on how I might optimize this query?
As always, it's all about the indexes. The index it chose on your
main table looks pretty weak. You probably should move those
non-joining columns out of your join condition in
- Original Message -
> From: "Daevid Vincent"
>
> d.`date_release`!='-00-00' AND sd.`scene_id`=s.`scene_id` GROUP
> BY sg.`scene_id` ORDER BY num DESC, sd.`count_dl_monthly` DESC LIMIT 30;
Could you put an index on sg.scene_id? Not sure, but it might help the grouping
be more effici
> Is it ALWAYS possible to fabricate a query/schema in
> such a way that MySQL ALWAYS uses the ideal
No. Optimisation is better in 5.6 than in 5.0, though.
Did you try adding multi-column indexes to cover the join and the order
by clause?
> 'Using where' extra -- you just have to keep at it?
s 2011/05/27 12:26 -0700, Daevid Vincent
[a] the EXTENDED keyword doesn't seem to do anything different? I get the
same columns and results??!
"show warnings"
2011/05/27 12:26 -0700, Daevid Vincent
In other words, am I wasting my time trying to tweak my
query and indexe
