To: Mitch Pirtle <mitch.pirtle@gmail.com>
Cc: "pgsql-performance@postgresql.org" <pgsql-performance@postgresql.org>
Subject: Re: Joining views disables indexes? 
In-reply-to: <330532b60511011516u3709d826w391a4ac167c6967b@mail.gmail.com> 
References: <330532b60511011516u3709d826w391a4ac167c6967b@mail.gmail.com>
Comments: In-reply-to Mitch Pirtle <mitch.pirtle@gmail.com>
	message dated "Tue, 01 Nov 2005 18:16:59 -0500"
Date: Tue, 01 Nov 2005 18:28:44 -0500
Message-ID: <29380.1130887724@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

Mitch Pirtle <mitch.pirtle@gmail.com> writes:
> I have a client that is testing an internal data platform, and they
> were happy with PostgreSQL until they tried to join views - at that
> time they discovered PostgreSQL was not using the indexes, and the
> queries took 24 hours to execute as a result.

You'll need to provide some actual details if you want useful comments.
Let's see the table schemas, the view definitions, and the EXPLAIN plan
(I'll spare you a request for EXPLAIN ANALYZE given that it'd take 24
hours to get ;-) ... although some estimate of the number of rows
expected would be helpful).  And I trust they remembered to ANALYZE the
underlying tables first?  Also, which PG version exactly?

			regards, tom lane