From: Ron Mayer <rm_pg@cheapcomplexdevices.com>
Subject: Re: half the query time in an unnecessary(?) sort?
Date: Mon, 25 Apr 2005 22:45:57 -0700
Organization: Hub.Org Networking Services
Lines: 24
Message-ID: <426DD595.1040608@cheapcomplexdevices.com>
References: <426D4FA1.1050000@cheapcomplexdevices.com>
	<200504251511.41949.josh@agliodbs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
To: Josh Berkus <josh@agliodbs.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
In-Reply-To: <200504251511.41949.josh@agliodbs.com>
To: pgsql-performance@postgresql.org


Josh Berkus wrote: [quoted out of order]
>Ron,
> 
> Looking at your analyze, though, I think it's not the sort that's taking the 
> time as it is that the full sorted entity_id column won't fit in work_mem.  
> Try increasing it?

Yup, that indeed fixed this particular query since neither table was
particularly large.


> It still has to sort because the clustering isn't guarenteed to be 100%.   

I guess I was contemplating whether or not there are some conditions
where it could be 100% (perhaps combined with Hannu's read only
table speculation).

> However, such sorts should be very quick as they have little work to do.

True, so long as the table can fit in work-mem.   For much larger tables
IMHO it'd be nice to be able to simply do a seq-scan on them if there were
some way of knowing that they were sorted.