Date: Thu, 22 Feb 2007 08:56:48 -0800
From: David Fetter <david@fetter.org>
To: Sergio Andreozzi <sergio.andreozzi@gmail.com>
Cc: pgsql-general@postgresql.org
Subject: Re: how to generate a list of distinct scalar values from a column
	which type is array
Message-ID: <20070222165648.GE10321@fetter.org>
References: <2cfc7c940702211021u1ba47031l25112a69cfb18338@mail.gmail.com>
	<2cfc7c940702220338h3eb4ec97p8cd759cb7a36d775@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <2cfc7c940702220338h3eb4ec97p8cd759cb7a36d775@mail.gmail.com>
User-Agent: Mutt/1.4.2.1i

On Thu, Feb 22, 2007 at 12:38:50PM +0100, Sergio Andreozzi wrote:
> Dear all,
> 
> given a column which type is for instance varchar(20)[],

This is almost never a good design.  If you must have an interface
like that, make it VIEW over an aggregate, which you can make
writeable.

> is it possible via SQL to generate the list of distinct scalar
> values?

Yes, but you should fix your design :)

CREATE TABLE foo (ft TEXT[]);

COPY foo(ft) FROM stdin;
{aaa,bb,c}
{dddd,eeee}
{aaa,eeee}
\.

SELECT DISTINCT ft[i]
FROM (
    SELECT ft,
        generate_series(
            array_lower(ft,1),
            array_upper(ft,1)
        ) AS i
    FROM foo
) AS bar;

So, yes, you can do it, and no, you shouldn't.

Cheers,
D
-- 
David Fetter <david@fetter.org> http://fetter.org/
phone: +1 415 235 3778        AIM: dfetter666
                              Skype: davidfetter

Remember to vote!