To: Joe Conway <mail@joeconway.com>
Cc: Richard Huxton <dev@archonet.com>, Andreas Tille <tillea@rki.de>,
	PostgreSQL SQL <pgsql-sql@postgresql.org>
Subject: Re: Scaler forms as function arguments
References: <Pine.LNX.4.56.0311261255210.17774@wr-linux02.rki.ivbb.bund.de>
	<200311261725.21975.dev@archonet.com> <3FC6E15E.2030300@joeconway.com>
In-Reply-To: <3FC6E15E.2030300@joeconway.com>
From: Greg Stark <gsstark@mit.edu>
Organization: The Emacs Conspiracy; member since 1992
Date: 28 Nov 2003 11:48:04 -0500
Message-ID: <87llq0fni3.fsf@stark.dyndns.tv>
Lines: 26
User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii


Joe Conway <mail@joeconway.com> writes:

> In 7.4 you could use an array. It would look like this:


Though note that 7.4 doesn't know how to optimize this form:


db=> explain select * from foo where foo_id in (1,2);
                                                   QUERY PLAN                                                    
-----------------------------------------------------------------------------------------------------------------
 Index Scan using foo_pkey, foo_pkey on foo  (cost=0.00..6.05 rows=2 width=756)
   Index Cond: ((foo_id = 1) OR (foo_id = 2))
(2 rows)


db=> explain select * from foo where foo_id = ANY (array[1,2]);
                              QUERY PLAN                              
----------------------------------------------------------------------
 Seq Scan on foo  (cost=0.00..1132.82 rows=5955 width=756)
   Filter: (foo_id = ANY ('{1,2}'::integer[]))
(2 rows)

-- 
greg