Date: Fri, 30 Mar 2007 04:55:34 -0600
From: Michael Fuhr <mike@fuhr.org>
To: Gaetano Mendola <mendola@bigfoot.com>
Cc: "pgsql-performance@postgresql.org" <pgsql-performance@postgresql.org>,
	Claus Guttesen <kometen@gmail.com>
Subject: Re: Wrong plan sequential scan instead of an index one
Message-ID: <20070330105534.GA1737@winnie.fuhr.org>
References: <460CD941.6080303@bigfoot.com>
	<b41c75520703300251oe916648r75c16c1bf70cd10@mail.gmail.com>
	<460CE19A.3060909@bigfoot.com>
	<20070330104611.GA1540@winnie.fuhr.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20070330104611.GA1540@winnie.fuhr.org>

On Fri, Mar 30, 2007 at 04:46:11AM -0600, Michael Fuhr wrote:
> Have you tried increasing the statistics target on l_pvcp.value?
> I ran your queries against canned data in 8.2.3 and better statistics
> resulted in more accurate row count estimates for this and other
> parts of the plan.  I don't recall if estimates for non-leading-character
> matches in earlier versions can benefit from better statistics.

This might work only in 8.2.  I see the following in the Release Notes:

* Improve the optimizer's selectivity estimates for LIKE, ILIKE,
  and regular expression operations (Tom)

-- 
Michael Fuhr