Received: from localhost (unknown [200.46.204.182])
	by postgresql.org (Postfix) with ESMTP id 10DA52E2DE4
	for <pgsql-sql-postgresql.org@postgresql.org>;
	Tue, 22 Jan 2008 08:58:15 -0400 (AST)
Received: from postgresql.org ([200.46.204.71])
	by localhost (mx1.hub.org [200.46.204.182]) (amavisd-maia, port 10024)
	with ESMTP id 49126-03-2 for <pgsql-sql-postgresql.org@postgresql.org>;
	Tue, 22 Jan 2008 08:58:07 -0400 (AST)
Received: from cat.co.za (mailer.cat.co.za [196.34.125.51])
	by postgresql.org (Postfix) with SMTP id 40A622E38FD
	for <pgsql-sql@postgresql.org>; Tue, 22 Jan 2008 08:55:28 -0400 (AST)
Received: (qmail 21738 invoked by uid 1002); 22 Jan 2008 12:55:53 -0000
Received: from stuartb@cat.co.za by mail.cat.co.za by uid 807 with
	qmail-scanner-1.21 (clamscan: 0.70.  Clear:RC:1(196.34.125.52):.
	Processed in 3.509096 secs); 22 Jan 2008 12:55:53 -0000
Received: from unknown (HELO ?192.168.34.201?) (196.34.125.52)
	by 0 with SMTP; 22 Jan 2008 12:55:49 -0000
Message-ID: <4795E6FD.4080107@cat.co.za>
Date: Tue, 22 Jan 2008 14:52:13 +0200
From: Stuart Brooks <stuartb@cat.co.za>
User-Agent: Thunderbird 2.0.0.9 (Windows/20071031)
MIME-Version: 1.0
To: Michael Glaesemann <grzm@seespotcode.net>
CC: pgsql-sql@postgresql.org
Subject: Re: TIMESTAMP comparison problem
References: <4795A3B0.3010505@cat.co.za>
	<794E6600-42A8-41A2-822D-080E19F9F0CC@seespotcode.net>
In-Reply-To: <794E6600-42A8-41A2-822D-080E19F9F0CC@seespotcode.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: Maia Mailguard 1.0.1
X-Archive-Number: 200801/175
X-Sequence-Number: 30028


>> If I reduce the precision to 3 for the timestamps it appears to work 
>> although it makes me nervous.
>
> With float timestamps, you're fooling yourself if you think those 
> numbers past the decimal are reliable.
>
>> Should I just recompile to use integer datetimes? I would like to 
>> have at least microsecond precision.
>
> Well, you can't get better than microsecond precision with timestamps 
> in Postgres. And the only way you can rely on that level of precision 
> is to compile with --enable-integer-datetimes.
>
> Michael Glaesemann 
I thought that might be the case, thanks for the help,

Stuart