Subject: Re: Performance considerations for very heavy INSERT
From: Ian Westmacott <ianw@intellivid.com>
To: blblack@gmail.com
Cc: pgsql-performance@postgresql.org
In-Reply-To: <84621a6050913083079ef0559@mail.gmail.com>
References: <84621a6050912140464c380c3@mail.gmail.com>
	<59d991c40509122039773cac17@mail.gmail.com>
	<84621a6050913083079ef0559@mail.gmail.com>
Content-Type: text/plain
Organization: Intellivid Corp.
Message-Id: <1126628214.11675.55.camel@spectre.intellivid.com>
Mime-Version: 1.0
Date: Tue, 13 Sep 2005 12:16:55 -0400
Content-Transfer-Encoding: 7bit

On Tue, 2005-09-13 at 11:30, Brandon Black wrote:
> I started looking closer at my options for partitioning (inheritance,
> union all), and at Bizgres today.  Bizgres partitioning appears to be
> basically the same kind of inheritance partitioning one can do in
> mainline PostgreSQL.  Am I correct in thinking that the main
> difference is that they've coded support for
> "enable_constraint_exclusion=true" so that the query planner can be
> more effective at taking advantage of the partitioning when you've
> specified CHECK constraints on the child tables?  I may go for 8.1
> instead in that case, as the main win I'm looking for is that with
> inheritance I'll be doing inserts into smaller tables instead of ones
> that grow to unmanageable sizes (and that I can drop old child tables
> instead of delete/vacuum).

Perhaps I missed something in this thread, but don't forget
you still need vacuum to reclaim XIDs.

	--Ian