Received: from localhost (unknown [200.46.204.184]) by postgresql.org (Postfix) with ESMTP id 344B72E1D81 for ; Wed, 14 Nov 2007 08:54:12 -0400 (AST) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.184]) (amavisd-maia, port 10024) with ESMTP id 07965-03 for ; Wed, 14 Nov 2007 08:53:43 -0400 (AST) Received: from bonito.ulb.ac.be (bonito.ulb.ac.be [164.15.59.220]) by postgresql.org (Postfix) with ESMTP id 63BB12E0F8F for ; Wed, 14 Nov 2007 08:54:08 -0400 (AST) Received: from [10.0.0.194] (bebif01.ulb.ac.be [164.15.125.4]) by bonito.ulb.ac.be (Postfix) with ESMTP id DCE633EC; Wed, 14 Nov 2007 13:54:06 +0100 (CET) Subject: Re: ALL() question From: Julien Cigar To: Richard Huxton Cc: pgsql-sql@postgresql.org In-Reply-To: <1195051800.3190.18.camel@frodon.be-bif.ulb.ac.be> References: <1195047542.3190.9.camel@frodon.be-bif.ulb.ac.be> <473AE27E.6050103@archonet.com> <1195051800.3190.18.camel@frodon.be-bif.ulb.ac.be> Content-Type: text/plain Date: Wed, 14 Nov 2007 15:57:50 +0100 Message-Id: <1195052270.3190.21.camel@frodon.be-bif.ulb.ac.be> Mime-Version: 1.0 X-Mailer: Evolution 2.12.0 Content-Transfer-Encoding: 7bit X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0.001 tagged_above=0 required=5 tests=BAYES_50=0.001 X-Spam-Level: X-Archive-Number: 200711/65 X-Sequence-Number: 29593 I finally found a solution: SELECT specimen_id FROM specimen_test_bits GROUP BY specimen_id HAVING array_accum(test_bit_id) = '{2,3,4}'; .. but I don't think it's very "clean" .. what do you think ? Thanks On Wed, 2007-11-14 at 15:50 +0100, Julien Cigar wrote: > On Wed, 2007-11-14 at 11:56 +0000, Richard Huxton wrote: > > Julien Cigar wrote: > > > > > > What I would like is a query that returns all the specimen_id of > > > this table which have _all_ the given test_bit_id. > > [snip] > > > With the following I got a syntax error: > > > select specimen_id > > > from specimen_test_bits > > > where test_bit_id = all(1,2,3,4); > > > > It's expecting an array here. You'd have to write > > = all('{1,2,3,4}') > > But that would have the same problem as... > > > > > The following works but no rows are returned : > > > select specimen_id > > > from specimen_test_bits > > > where test_bit_id = all(select id from test_bits where id in (1,2,3,4)); > > > > It's testing each row individually and of course one row can't match ALL > > four values. > > > > What you want to do is count the distinct values. Something like: > > > > SELECT > > specimen_id > > FROM foo > > GROUP BY > > specimen_id > > HAVING > > count(distinct test_bit_id) = 4 > > ; > > > > I don't think it would work, for example if I have: > specimen_id | test_bit_id > ------------+------------ > 100 1 > 100 3 > 101 1 > 101 2 > > the test_bit_ids are parameters, so with the given test_bit_id 1,3 it > would return specimen_id 101 too, which I don't want ... > What I would like is the specimen_id which match _exactly_ the given > test_bit_ids, so it should return only 100 in this example .. > > from the documentation ALL() can take a subquery too, not only an ARRAY > (http://www.postgresql.org/docs/8.2/static/functions-subquery.html) > >