Subject: Re: Delete query takes exorbitant amount of time
From: Simon Riggs <simon@2ndquadrant.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Stephan Szabo <sszabo@megazone.bigpanda.com>,
	Karim A Nassar <Karim.Nassar@NAU.EDU>,
	Christopher Kings-Lynne <chriskl@familyhealth.com.au>,
	pgsql-performance@postgresql.org
In-Reply-To: <21616.1112108201@sss.pgh.pa.us>
References: <Pine.SOL.4.21.0503290120550.5642-100000@coruscant.cet.nau.edu>
	<1112095776.11750.958.camel@localhost.localdomain>
	<20050329054714.V50612@megazone.bigpanda.com>
	<20050329061837.A51620@megazone.bigpanda.com>
	<21616.1112108201@sss.pgh.pa.us>
Content-Type: text/plain
Organization: 2nd Quadrant
Date: Tue, 29 Mar 2005 16:24:40 +0100
Message-Id: <1112109880.11750.977.camel@localhost.localdomain>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit

On Tue, 2005-03-29 at 09:56 -0500, Tom Lane wrote:
> Stephan Szabo <sszabo@megazone.bigpanda.com> writes:
> > If there were some way to pass a "limit" into SPI_prepare that was treated
> > similarly to a LIMIT clause for planning purposes but didn't actually
> > change the output plan to only return that number of rows, we could use
> > that.
> 
> Hmm ... the planner does have the ability to do that sort of thing (we
> use it for cursors).  SPI_prepare doesn't expose the capability.
> Perhaps adding a SPI_prepare variant that does expose it would be the
> quickest route to a solution.
> 
> I get a headache every time I look at the RI triggers ;-).  Do they
> always know at the time of preparing a plan which way it will be used?

If action is NO ACTION or RESTRICT then 
	we need to SELECT at most 1 row that matches the criteria
	which means we can use LIMIT 1

If action is CASCADE, SET NULL, SET DEFAULT then
	we need to UPDATE or DELETE all rows that match the criteria
	which means we musnt use LIMIT and need to use FOR UPDATE

We know that at CONSTRAINT creation time, which always occurs before
plan preparation time.

Best Regards, Simon Riggs