Date: Mon, 30 May 2005 23:08:01 -0500
From: Bruno Wolff III <bruno@wolff.to>
To: Tobias Brox <tobias@nordicbet.com>
Cc: pgsql-performance@postgresql.org
Subject: Re: Index on a NULL-value
Message-ID: <20050531040801.GA21334@wolff.to>
Mail-Followup-To: Bruno Wolff III <bruno@wolff.to>,
	Tobias Brox <tobias@nordicbet.com>, pgsql-performance@postgresql.org
References: <20050531030207.GD26100@tobias.exoweb.net>
	<20050531031653.GA17723@wolff.to>
	<20050531033158.GI26100@tobias.exoweb.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20050531033158.GI26100@tobias.exoweb.net>
User-Agent: Mutt/1.5.6i

On Tue, May 31, 2005 at 11:31:58 +0800,
  Tobias Brox <tobias@nordicbet.com> wrote:
> [Tobias Brox]
> > test=# set enable_seqscan=off; 
> 
> [Bruno Wolff III - Mon at 10:16:53PM -0500]
> > It isn't surprising that an index wasn't used since a sequential scan is
> > going to be faster in your test case.
> > 
> > If you want to test this out, you to want use realistically sized tables.
> 
> Wrong.  In this case I was not wondering about the planners choise of not
> using the index, but the fact that the planner could not find the index at
> all.  Reproducing it on a simple table in a test environment was a valid
> strategy to solve this specific problem.

I missed that you turned sequential scans off for your test.