Date: Sat, 20 Aug 2005 17:10:25 +0200
From: "Steinar H. Gunderson" <sgunderson@bigfoot.com>
To: pgsql-performance@postgresql.org
Subject: Re: index as large as table
Message-ID: <20050820151025.GA12890@uio.no>
Mail-Followup-To: pgsql-performance@postgresql.org
References: <Pine.UW2.4.53.0508201413430.7116@sun.pyrenet>
	<Pine.LNX.4.58.0508202302270.4243@linuxworld.com.au>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.58.0508202302270.4243@linuxworld.com.au>
User-Agent: Mutt/1.5.10i

On Sat, Aug 20, 2005 at 11:08:13PM +1000, Gavin Sherry wrote:
> Of course. The idea is that, generally speaking, you're only interested in
> a small portion of the data stored in the table. Indexes store extra data
> so that they can locate the portion you're interested in faster.

I think his question was more why you needed the data in itself, when you had
everything you needed in the index anyway. (Actually, you don't -- indexes
don't carry MVCC information, but I guess that's a bit beside the point.)

There has been discussion on "heap tables" or whatever you'd want to call
them (ie. tables that are organized as a B+-tree on some index) here before;
I guess the archives would be a reasonable place to start looking.

/* Steinar */
-- 
Homepage: http://www.sesse.net/