Content-Type: text/plain;
  charset="iso-8859-1"
From: Josh Berkus <josh@agliodbs.com>
Organization: Aglio Database Solutions
To: dev@archonet.com, val@webtribe.net
Subject: Re: which will be faster? w/ or w/o indices
Date: Fri, 7 Feb 2003 09:26:22 -0800
User-Agent: KMail/1.4.3
Cc: pgsql-sql@postgresql.org
References: <20030207080420.2131718332@cheetah.webtribe.net>
	<1064.192.168.1.16.1044610614.squirrel@mainbox.archonet.com>
In-Reply-To: <1064.192.168.1.16.1044610614.squirrel@mainbox.archonet.com>
MIME-Version: 1.0
Message-Id: <200302070926.22649.josh@agliodbs.com>
Content-Transfer-Encoding: quoted-printable

Val,

> Can't really say without testing - it'll depend on the size of the various
> tables, X and Y disk speeds etc. The only suggestion I can make is to try
> an EXPLAIN (Reference manual, SQL commands) with and without
> ENABLE_SEQSCAN off (Admin guide ch 3.2).

Please use EXPLAIN ANALYZE instead of just EXPLAIN.    That will show you=
=20
whether you have performance problems due to things outside of the planner'=
s=20
control.

Keep in mind, though, that if your system has no special problems and your=
=20
planner statistics are up to date, the planner will do a good job of decidi=
ng=20
whether or not indexes are faster.   You don't need to start out=20
second-guessing the planner ... you're likely to be wrong more often that i=
t=20
is.   Save that for real problem queries.

Also, upgrade to 7.2.4!!!    7.2.1 has several significant bugs, including =
two=20
security exploits and one backup-and-restore bug that could make it difficu=
lt=20
to recover your database from a backup file.

--=20
Josh Berkus
Aglio Database Solutions
San Francisco