Received: from localhost (unknown [200.46.204.183])
	by mail.postgresql.org (Postfix) with ESMTP id 32E4D63225B
	for <pgsql-performance-postgresql.org@mail.postgresql.org>;
	Tue,  6 Jan 2009 09:22:08 -0400 (AST)
Received: from mail.postgresql.org ([200.46.204.86])
	by localhost (mx1.hub.org [200.46.204.183]) (amavisd-maia, port 10024)
	with ESMTP id 54287-08
	for <pgsql-performance-postgresql.org@mail.postgresql.org>;
	Tue,  6 Jan 2009 09:22:03 -0400 (AST)
X-Greylist: from auto-whitelisted by SQLgrey-1.7.6
Received: from sss.pgh.pa.us (sss.pgh.pa.us [66.207.139.130])
	by mail.postgresql.org (Postfix) with ESMTP id B114863223D
	for <pgsql-performance@postgresql.org>;
	Tue,  6 Jan 2009 09:22:03 -0400 (AST)
Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1])
	by sss.pgh.pa.us (8.14.2/8.14.2) with ESMTP id n06DM0hL007833;
	Tue, 6 Jan 2009 08:22:00 -0500 (EST)
To: "Robert Haas" <robertmhaas@gmail.com>
cc: "pgsql-performance@postgresql.org" <pgsql-performance@postgresql.org>
Subject: Re: bad selectivity estimates for CASE 
In-reply-to: <603c8f070901060425t485fb8d8nb79735ed72982d5a@mail.gmail.com> 
References: <603c8f070901051915p15b3d6a7jbe6f6fdd178143e@mail.gmail.com>
	<3032.1231216843@sss.pgh.pa.us>
	<603c8f070901060425t485fb8d8nb79735ed72982d5a@mail.gmail.com>
Comments: In-reply-to "Robert Haas" <robertmhaas@gmail.com>
	message dated "Tue, 06 Jan 2009 07:25:29 -0500"
Date: Tue, 06 Jan 2009 08:22:00 -0500
Message-ID: <7832.1231248120@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>
X-Virus-Scanned: Maia Mailguard 1.0.1
X-Spam-Status: No, hits=0 tagged_above=0 required=5 tests=none
X-Spam-Level: 
X-Archive-Number: 200901/11
X-Sequence-Number: 32290

"Robert Haas" <robertmhaas@gmail.com> writes:
> Well, presumably CASE WHEN <expr1> THEN <constant1> WHEN <expr2> THEN
> <constant2> WHEN <expr3> THEN <constant3> ... END = <constantn> could
> be simplified to <exprn>.

Not without breaking the order-of-evaluation guarantees.  Consider

	case when x=0 then 0 when 1/x = 42 then 1 end = 1

This expression should not suffer a divide-by-zero failure but your
proposal would allow it to do so.

			regards, tom lane