To: pgsql-sql@postgresql.org
Subject: Selecting exactly one row for each column value
From: Florian Weimer <fweimer@bfk.de>
Date: Tue, 06 Mar 2007 16:03:36 +0100
Message-ID: <82hcsy4do7.fsf@mid.bfk.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

I've got the following table:

fweimer=3D> SELECT * FROM tab;
 a | b | c
---+---+---
 1 | 2 | 3
 5 | 6 | 7
 1 | 2 | 2
 2 | 3 | 4
 1 | 2 | 2
 2 | 3 | 4

For each value in the first column, I need one (and only one) matching
row from the table.  A possible solution is:

 a | b | c
---+---+---
 5 | 6 | 7
 2 | 3 | 4
 1 | 2 | 3

Of course,

SELECT a, (SELECT b FROM tab  i  WHERE  i.a =3D o.a LIMIT 1),=20
     (SELECT c FROM TAB i WHERE i.a =3D o.a LIMIT 1) FROM tab o GROUP BY o.=
a;

does the trick, but this approach seems to rely on undefined behavior
and quickly gets messy when the number of columns increases.

Is there a better way to implement this?

--=20
Florian Weimer                <fweimer@bfk.de>
BFK edv-consulting GmbH       http://www.bfk.de/
Kriegsstra=DFe 100              tel: +49-721-96201-1
D-76133 Karlsruhe             fax: +49-721-96201-99