Skip site navigation (1) Skip section navigation (2)

Re: Column correlation drifts, index ignored again

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Kevin Brown <kevin(at)sysexperts(dot)com>
Cc: Postgres Performance <pgsql-performance(at)postgresql(dot)org>
Subject: Re: Column correlation drifts, index ignored again
Date: 2004-02-24 22:16:08
Message-ID: 27171.1077660968@sss.pgh.pa.us (view raw or flat)
Thread:
Lists: pgsql-performance
Kevin Brown <kevin(at)sysexperts(dot)com> writes:
> One problem I've been running into is the merge join spilling to disk
> because sort_mem isn't big enough.  The problem isn't that this is
> happening, it's that I think the planner is underestimating the impact
> that doing this will have on the time the merge join takes.  Does the
> planner even account for the possibility that a sort or join will spill
> to disk?

Yes it does.  I thought it was making a pretty good estimate, actually.
The only obvious hole in the assumptions is

 * The disk traffic is assumed to be half sequential and half random
 * accesses (XXX can't we refine that guess?)

Because of the way that tuplesort.c works, the first merge pass should
be pretty well sequential, but I think the passes after that might be
mostly random from the kernel's viewpoint :-(.  Possibly the I/O cost
should be adjusted depending on how many merge passes we expect.


> In any case, one thing that none of this really accounts for is that
> it's better to set random_page_cost too low than too high.

That depends on what you are doing, although I will concede that a lot
of people are doing things where indexscans should be favored.

			regards, tom lane

In response to

pgsql-performance by date

Next:From: Jonathan GardnerDate: 2004-02-24 22:19:39
Subject: Re: [HACKERS] [SQL] Materialized View Summary
Previous:From: Robert TreatDate: 2004-02-24 21:48:49
Subject: Re: [HACKERS] [SQL] Materialized View Summary

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group