Subject: Re: BETWEEN optimizer problems with single-value
From: Simon Riggs <simon@2ndquadrant.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Kevin Grittner <Kevin.Grittner@wicourts.gov>,
	pgsql-performance@postgresql.org,
	Andreas Kretschmer <akretschmer@spamfence.net>
In-Reply-To: <1142467626.3859.207.camel@localhost.localdomain>
References: <44180105.EE98.0025.0@wicourts.gov>
	<20060315181735.GA22240@KanotixBox> <44180D33.EE98.0025.0@wicourts.gov>
	<28798.1142450270@sss.pgh.pa.us>
	<1142467626.3859.207.camel@localhost.localdomain>
Content-Type: text/plain
Date: Thu, 16 Mar 2006 00:24:38 +0000
Message-Id: <1142468678.3859.221.camel@localhost.localdomain>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit

On Thu, 2006-03-16 at 00:07 +0000, Simon Riggs wrote:
> On Wed, 2006-03-15 at 14:17 -0500, Tom Lane wrote:
> 
> > It looks to me like this is a matter of bad cost estimation, ie, it's
> > thinking the other index is cheaper to use.  Why that is is not clear.
> > Can we see the pg_stats rows for ctofcNo and calDate?
> 
> ISTM that when the BETWEEN constants match we end up in this part of
> clauselist_selectivity()...

(and now for the whole email...)

	/*
	 * It's just roundoff error; use a small positive
	 * value
	 */
	s2 = 1.0e-10;

so that the planner underestimates the cost of using "Cal_CalDate" so
that it ends up the same as "Cal_CtofcNo", and then we pick
"Cal_CalDate" because it was created first.

Using 1.0e-10 isn't very useful... the selectivity for a range should
never be less than the selectivity for an equality, so we should simply
put in a test against one of the pseudo constants and use that as the
minimal value. That should lead to raising the apparent cost of
Cal_CalDate so that Cal_CtofcNo can take precedence.

Best Regards, Simon Riggs