Message-ID: <4506B9E2.9050608@enterprisedb.com>
Date: Tue, 12 Sep 2006 14:45:06 +0100
From: Heikki Linnakangas <heikki@enterprisedb.com>
User-Agent: Thunderbird 1.5.0.5 (X11/20060812)
MIME-Version: 1.0
To: alvherre@commandprompt.com, Guillaume Cottenceau <gc@mnc.ch>,
	pgsql-performance@postgresql.org
Subject: Re: Poor performance on seq scan
References: <450684EC.4090304@designaproduct.biz>
	<4506902C.7090407@enterprisedb.com>
	<4506A9E7.2060303@designaproduct.biz>
	<87y7spw8v9.fsf@meuh.mnc.lan> <4506B508.80806@enterprisedb.com>
	<20060912133254.GC11324@alvh.no-ip.org>
In-Reply-To: <20060912133254.GC11324@alvh.no-ip.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Alvaro Herrera wrote:
> Are you saying that an indexscan "Filter" only acts after getting the
> heap tuple? If that's the case, then there's room for optimization
> here, namely if the affected column is part of the index key, then we
> could do the filtering before fetching the heap tuple.

That's right. Yes, there's definitely room for optimization. In general, 
it seems we should detach the index scan and heap fetch more. Perhaps 
make them two different nodes, like the bitmap index scan and bitmap 
heap scan. It would allow us to do the above. It's also going to be 
necessary if we ever get to implement index-only scans.

-- 
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com