Re: Forcing postgresql to use an index - Mailing list pgsql-performance

From Eugene Morozov
Subject Re: Forcing postgresql to use an index
Date
Msg-id 87eiqh77fq.fsf@eugenemorozov.name
Whole thread Raw
In response to Forcing postgresql to use an index  (Eugene Morozov <eugene@cactus-mouse.com>)
List pgsql-performance
Grzegorz Jaśkiewicz <gryzman@gmail.com> writes:

> Learn it to not generate with "WITH IN (subq)", is this can be quite
> slow on postgresql. Use joins instead.

OK, I've split the query in two (can't make Django to generate JOIN in this
case) and it always uses index now. This immediately opened road for
other optimizations. Thanks!

>
> looks like planner was wrong about rowcount in one place: Hash IN Join
>  (cost=2204.80..4809.31 rows=292 width=202) (actual
> time=12.856..283.916 rows=15702 loops=1)
>
> I have no idea why, probably more knowledgeable guys will know more
> about why. But overall, all other stats seem to be okay.
> What's the default_statistics_target setting in the postgresql set to?
>
> One thing tho, what's the version, and platform.

PostgreSQL 8.3.7, Ubuntu 8.10

pgsql-performance by date:

Previous
From: Vratislav Benes
Date:
Subject: PSQL 8.4 - partittions - join tables - not optimal plan
Next
From: "A. Kretschmer"
Date:
Subject: Re: PSQL 8.4 - partittions - join tables - not optimal plan