Message-ID: <3DF90A89.61040846@nsd.ca>
Date: Thu, 12 Dec 2002 17:15:37 -0500
From: Jean-Luc Lachance <jllachan@nsd.ca>
MIME-Version: 1.0
To: Stephan Szabo <sszabo@megazone23.bigpanda.com>
Cc: pgsql-general@postgresql.org, pgsql-performance@postgresql.org
Subject: Re: CLUSTER command
References: <20021212135913.Q11714-100000@megazone23.bigpanda.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

The current cluster command is equivalant to:

create b as select * from a order by i;

So you would not be loosing anything.


Stephan Szabo wrote:
> 
> On Thu, 12 Dec 2002, Jean-Luc Lachance wrote:
> 
> > Hi all,
> >
> > I just read about the cluster command and was a little (very)
> > disapointed.
> > Clustered tables do not remain clustered after inserts.
> > Clustered tables are usefull when the table is very large and there are
> > few different keys.
> >
> >
> > Because the table file is already extended (2G limit) using different
> > files extension (.N)
> > how complicated (modifying the code) would it be to have the table files
> > split according to the cluster key?
> 
> I'd vote against changing the existing CLUSTER since the existing CLUSTER
> while not great does handle many different key values fairly well as well
> and this solution wouldn't.  Many different key values are still
> useful to cluster if you're doing searches over ranges since it lowers the
> number of heap file reads necessary.  If done this should probably be
> separate from the existing cluster or at least both versions should be
> possible.