From: Gregory Stark <stark@enterprisedb.com>
To: "Alvaro Herrera" <alvherre@commandprompt.com>
Cc: <pgsql-performance@postgresql.org>
Subject: Re: What's The Difference Between VACUUM and VACUUM ANALYZE?
In-Reply-To: <20070508222102.GD24167@alvh.no-ip.org> (Alvaro Herrera's message
	of "Tue, 8 May 2007 18:21:02 -0400")
Organization: EnterpriseDB
References: <b09064f30705081443w25bf2e6o3a663c5f3a2a7798@mail.gmail.com>
	<20070508215213.GC24167@alvh.no-ip.org> <20070508220608.GA8457@uio.no>
	<20070508222102.GD24167@alvh.no-ip.org>
Date: Wed, 09 May 2007 12:31:11 +0100
Message-ID: <87hcqmi680.fsf@oxford.xeocode.com>
User-Agent: Gnus/5.110006 (No Gnus v0.6) Emacs/21.4 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

"Alvaro Herrera" <alvherre@commandprompt.com> writes:

> Steinar H. Gunderson wrote:
>
>> Shoot me if I'm wrong here, but doesn't VACUUM ANALYZE check _all_ tuples,
>> as compared to the random selection employed by ANALYZE?
>
> You are wrong, but it won't be me the one to shoot you.
>
> There have been noises towards making the ANALYZE portion use the same
> scan that VACUUM already does, but nobody has written the code (it would
> be useful for some kinds of stats).

I think it does for the count of total records in the table. 
But not for the rest of the stats.

-- 
  Gregory Stark
  EnterpriseDB          http://www.enterprisedb.com