Date: Thu, 15 Dec 2005 21:48:55 -0800
From: Kevin Brown <kevin@sysexperts.com>
To: pgsql-performance@postgresql.org
Subject: Re: Overriding the optimizer
Message-ID: <20051216054855.GD691@filer>
Mail-Followup-To: Kevin Brown <kevin@sysexperts.com>,
	pgsql-performance@postgresql.org
References: <7484.1134600423@sss.pgh.pa.us>
	<1134629567.14248.27.camel@firebolt>
	<200512150148.16125.blargity@gmail.com>
	<1134644526.14248.60.camel@firebolt>
	<43A1C40D.8040006@paradise.net.nz>
	<43A1F6DB.3080805@modgraph-usa.com>
	<43A226E7.7040508@paradise.net.nz>
	<43A22BE3.6020103@modgraph-usa.com> <20051216052025.GB691@filer>
	<43A25372.8090706@modgraph-usa.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
In-Reply-To: <43A25372.8090706@modgraph-usa.com>
Organization: Frobozzco International
User-Agent: Mutt/1.5.11

Craig A. James wrote:
> Kevin Brown wrote:
> >>Hints are dangerous, and I consider them a last resort.
> >
> >If you consider them a last resort, then why do you consider them to
> >be a better alternative than a workaround such as turning off
> >enable_seqscan, when all the other tradeoffs are considered?
> 
> If I understand enable_seqscan, it's an all-or-nothing affair.  Turning it 
> off turns it off for the whole database, right?  The same is true of all 
> of the planner-tuning parameters in the postgres conf file.

Nope.  What's in the conf file are the defaults.  You can change them
on a per-connection basis, via the SET command.  Thus, before doing
your problematic query:

SET enable_seqscan = off;

and then, after your query is done, 

SET enable_seqscan = on;

> >If your argument is that planner hints would give you finer grained
> >control, then the question is whether you'd rather the developers
> >spend their time implementing planner hints or improving the planner.
> 
> I agree 100% -- I'd much prefer a better planner.  But when it comes down 
> to a do-or-die situation, you need a hack, some sort of workaround, to get 
> you working *today*.

And that's why I was asking about workarounds versus planner hints.  I
expect that the situations in which the planner gets things wrong
*and* where there's no workaround are very rare indeed.


-- 
Kevin Brown					      kevin@sysexperts.com