Message-ID: <4CB5F394.6090909@krogh.cc>
Date: Wed, 13 Oct 2010 19:59:48 +0200
From: Jesper Krogh <jesper@krogh.cc>
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US;
	rv:1.9.1.12) Gecko/20100915 Thunderbird/3.0.8
MIME-Version: 1.0
To: Robert Haas <robertmhaas@gmail.com>
CC: Neil Whelchel <neil.whelchel@gmail.com>, pgsql-performance@postgresql.org
Subject: Re: Slow count(*) again...
References: <201010091626.18636.neil.whelchel@gmail.com>
	<201010130138.39893.neil.whelchel@gmail.com>
	<4CB572CF.4080709@catalyst.net.nz>
	<201010130316.12922.neil.whelchel@gmail.com>
	<AANLkTi=Cn1N8keywMbePZPiezg9cn_EXQZiLzF7cym+q@mail.gmail.com>
In-Reply-To: <AANLkTi=Cn1N8keywMbePZPiezg9cn_EXQZiLzF7cym+q@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

On 2010-10-13 15:28, Robert Haas wrote:
> On Wed, Oct 13, 2010 at 6:16 AM, Neil Whelchel<neil.whelchel@gmail.com>  wrote:
>    
>> I might go as far as to rattle the cage of the developers to see if it makes
>> any sense to add some column oriented storage capability to Postgres. That
>> would be the hot ticket to be able to specify an attribute on a column so that
>> the back end could shadow or store a column in a column oriented table so
>> aggregate functions could work on them with good efficiency, or is that an
>> INDEX?
>>      
> I'd love to work on that, but without funding it's tough to find the
> time.  It's a big project.
>    
Is it hugely different from just getting the visibillity map suitable
for doing index-only scans and extracting values from the index
directly as Heikki has explained?

That would essentially do a column oriented table (the index itself)
of a specific columns (or column set).

... still a huge task though.

-- 
Jesper