From: Tom Lane <tgl@sss.pgh.pa.us>
To: Laszlo Nagy <gandalf@shopzeus.com>
cc: pgsql-admin@postgresql.org
Subject: Re: Messed up time zones
In-reply-to: <501BF70E.4090903@shopzeus.com>
References: <501B8C6F.6040100@shopzeus.com> <501BAA84.9010306@shopzeus.com>
	<501BAE33.4090102@shopzeus.com> <16399.1344003584@sss.pgh.pa.us>
	<501BED07.5070603@shopzeus.com> <501BEF7C.1030403@dhs-club.com>
	<501BF70E.4090903@shopzeus.com>
Comments: In-reply-to Laszlo Nagy <gandalf@shopzeus.com>
	message dated "Fri, 03 Aug 2012 18:06:38 +0200"
Date: Fri, 03 Aug 2012 12:38:11 -0400
Message-ID: <19252.1344011891@sss.pgh.pa.us>

Laszlo Nagy <gandalf@shopzeus.com> writes:
> I have intentionally chosen an example where the local time is changed 
> from summer time to winter time (e.g. local time suddenly "goes back" 
> one hour). It demonstrates that you cannot use "at time zone ...." 
> expression to convert a timestamptz into a desired time zone manually.

Um, yes you can.  The trick is to use a timezone name, not an
abbreviation, in the AT TIME ZONE construct (for instance,
'Europe/Budapest' not just 'CET').  That will do the rotation
in a DST-aware fashion.

> As far as I'm concerned, I'm going to set the system's clock to UTC, 
> store everything in timestamp field (in UTC),  and use a program to 
> convert fetched values before displaying them.

[ shrug... ]  If you really insist on re-inventing that wheel, go
ahead, but it sounds to me like you'll just be introducing additional
points of failure.

			regards, tom lane