X-Original-To: pgsql-sql-postgresql.org@localhost.postgresql.org
Received: from localhost (unknown [200.46.204.2])
	by svr1.postgresql.org (Postfix) with ESMTP id 322A7D1B4AC
	for <pgsql-sql-postgresql.org@localhost.postgresql.org>;
	Thu, 29 Jan 2004 18:28:33 +0000 (GMT)
Received: from svr1.postgresql.org ([200.46.204.71])
	by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024)
	with ESMTP id 52401-04
	for <pgsql-sql-postgresql.org@localhost.postgresql.org>;
	Thu, 29 Jan 2004 14:28:02 -0400 (AST)
Received: from megazone.bigpanda.com (megazone.bigpanda.com [64.147.171.210])
	by svr1.postgresql.org (Postfix) with ESMTP id 5B4CBD1B897
	for <pgsql-sql@postgresql.org>; Thu, 29 Jan 2004 14:28:01 -0400 (AST)
Received: by megazone.bigpanda.com (Postfix, from userid 1001)
	id 97D8E35794; Thu, 29 Jan 2004 10:27:57 -0800 (PST)
Received: from localhost (localhost [127.0.0.1])
	by megazone.bigpanda.com (Postfix) with ESMTP
	id 95FDF3578D; Thu, 29 Jan 2004 10:27:57 -0800 (PST)
Date: Thu, 29 Jan 2004 10:27:57 -0800 (PST)
From: Stephan Szabo <sszabo@megazone.bigpanda.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Bruno Wolff III <bruno@wolff.to>,
	Achilleus Mantzios <achill@matrix.gatewaynet.com>,
	email lists <lists@darrenmackay.com>, pgsql-sql@postgresql.org
Subject: Re: query not using index for descending records? 
In-Reply-To: <24411.1075397054@sss.pgh.pa.us>
Message-ID: <20040129101857.D26621@megazone.bigpanda.com>
References: <20040129131847.GE17068@wolff.to>
	<Pine.LNX.4.44.0401291525350.9854-100000@matrix.gatewaynet.com>
	<20040129140655.GA14885@wolff.to>
	<20040129075643.D23153@megazone.bigpanda.com>
	<24411.1075397054@sss.pgh.pa.us>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Virus-Scanned: by amavisd-new at postgresql.org
X-Archive-Number: 200401/271
X-Sequence-Number: 16664

On Thu, 29 Jan 2004, Tom Lane wrote:

> Stephan Szabo <sszabo@megazone.bigpanda.com> writes:
> > I'd thought that I'd previously sent a message containing a set of
> > definitions for the reverse opclasses (not meant for inclusion to the
> > system because I was making SQL functions that basically did -<normal
> > comparison function> to use as the function of the operator class, but
> > possibly worth playing with) but now that I actually search again, I don't
> > see it.
>
> I don't recall having seen such a thing go by...
>
> I fear that using SQL functions as comparators would only be useful for
> proof-of-concept, not as an industrial-strength implementation.  The
> index code needs comparator functions not to leak memory, and I doubt
> that that could be guaranteed with a SQL function.  You'd probably have
> speed issues too.

Yeah, that's what I figured.  I thought it might be useful for people to
play with though since at least for the integer/float types writing C
versions of the comparitors is easy. I was thinking for real it'd be nice
to be able to use the normal comparitor but invert the return value as
necessary rather than providing two functions, but I didn't look at what
that would take.