To: Christoph Haller <ch@rodos.fzk.de>
Cc: pgsql-sql@postgresql.org, cprice@hrdenterprises.com
Subject: Re: LIMIT and SUBQUERIES
References: <3E65FA8B.352136E@rodos.fzk.de>
In-Reply-To: <3E65FA8B.352136E@rodos.fzk.de>
From: Greg Stark <gsstark@mit.edu>
Organization: The Emacs Conspiracy; member since 1992
Date: 05 Mar 2003 10:22:46 -0500
Message-ID: <87fzq196t5.fsf@stark.dyndns.tv>
Lines: 25
User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

Christoph Haller <ch@rodos.fzk.de> writes:

> >
> > So anyway, I have the query that is working - but it returns all
> > records for all owners, when what I really want to do is return the
> > top 5 per each owner.
> >
> I've seen a lot of questions like this on the list before about
> filtering
> result sets. Within the most replies, people were told to use middleware
> 
> stuff like perl, awk, sed, ... to filter out unwanted rows, because SQL
> was not intended to do so.

This is the same problem as the "ranking" problem that I mentioned earlier.

http://archives.postgresql.org/pgsql-sql/2003-03/msg00013.php

Essentially we would need a kind of feature that has some similarities to user
defined aggregates but is not exactly the same thing. The feature doesn't
currently exist, and I haven't seen it in order databases, so I don't even
know exactly what it would look like.

-- 
greg