Date: Sat, 11 Mar 2006 22:42:12 -0700
From: Michael Fuhr <mike@fuhr.org>
To: Weimao Ke <wke@indiana.edu>
Cc: pgsql-sql@postgresql.org
Subject: Re: pgsql aggregate: conditional max
Message-ID: <20060312054212.GA25423@winnie.fuhr.org>
References: <4413AD1C.9050604@indiana.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4413AD1C.9050604@indiana.edu>

On Sun, Mar 12, 2006 at 12:09:48AM -0500, Weimao Ke wrote:
> I want to group by "aid" and choose the category (i.e., "cat") with the 
> largest "weight":
> 
> aid   |   max_weighted_cat
> ----+---------------------
> a1   |   Other
> a2   |   Drama
> a3   |   Adult

PostgreSQL has a non-standard DISTINCT ON clause that would work.
See the weather_reports example in the documentation for SELECT:

http://www.postgresql.org/docs/8.1/interactive/sql-select.html

Try this query against your example data:

SELECT DISTINCT ON (aid) aid, cat
FROM tablename
ORDER BY aid, weight DESC, cat;

If multiple rows for a given aid match that aid's max weight then
the above query will return the first matching row according to the
given sort order.

Some people object to DISTINCT ON because it's non-deterministic if
you don't order by enough columns.  Here's something more standard;
it'll return all rows that match a given aid's max weight:

SELECT aid, cat
FROM tablename AS t
JOIN (SELECT aid, max(weight) AS weight
      FROM tablename
      GROUP BY aid) AS s USING (aid, weight);

-- 
Michael Fuhr