Mail-Followup-To: pgsql-sql@postgresql.org
To: pgsql-sql@postgresql.org
From: Tim Landscheidt <tim@tim-landscheidt.de>
Subject: Re: grouping subsets
Date: Thu, 22 Jul 2010 11:31:23 +0000
Organization: <URI:http://www.tim-landscheidt.de/>
Lines: 38
Message-ID: <m3mxtju510.fsf@passepartout.tim-landscheidt.de>
References: <i28ubc$ie9$1@dough.gmane.org>
	<20100722100226.GE10348@a-kretschmer.de>
	<4C48230D.7030304@archonet.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Mail-Copies-To: never
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.1 (gnu/linux)
Cancel-Lock: sha1:JeUWHdoTnq/oKr1I7xIINNkPNdU=

Richard Huxton <dev@archonet.com> wrote:

>>> What I want to get is the values grouped by "subset", where a subset is a set of rows with identical column until the colum changes.
>>> Is there a way to get

>>> | 2 | B |
>>> | 4 | C |
>>> | 4 | B |
>>> | 3 | D |

>>> by SQL only?

>> I think, the problem is that there are 2 identical groups. I think, you
>> can write a pl/pgsql-proc, selecting all ordered by the date-field and
>> walking through the result to do the grouping, checking if the 2nd
>> column is different from the previous.

>> With plain SQL it's maybe possible too, but i don't know how ...

> It should be do-able in 8.4 onwards, look into windowing
> functions. In particular the lag() function:

> SELECT
>     mycode,
>     mydate,
>     lag(mycode) OVER (ORDER BY mydate) AS prev_code
> FROM
>     mytable
> ORDER BY mydate;

> It should be possible to use that as a subquery with an
> outer query that compares mycode=prev_code to get a run
> length.

Hmmm. Can the outer query be done without using "WITH
RECURSIVE"?

Tim