On Jul 6, 2011, at 9:24 AM, Alexey Klyukin wrote:
> So, I'd think there are 2 reasonable approaches to storing the
> timezone part:
> 
> 1. Store the timezone abbreviation (i.e. 'EST' along w/ the timestamp
> data). 
> 2. Assign OID to each of the timezones and store it w/ the timestamp.
> 
> The first option seem to avoid the necessity of creating a new system
> catalog for timezone information and the burden of updating it,
> because current implementation is already capable of translating
> abbreviations to useful timezone information. The question is, whether
> just a TZ abbreviation is sufficient to uniquely identify the timezone
> and get the offset and DST rules. If it's not sufficient, how
> conflicting TZ short names are handled in the current code (i.e. 'AT
> TIME ZONE ...')?

It's not enough. See the timezone_abbreviations setting.
--
Jim C. Nasby, Database Architect                   jim(at)nasby(dot)net
512.569.9461 (cell)                         http://jim.nasby.net

In response to

• Re: storing TZ along timestamps at 2011-07-06 14:24:52 from Alexey Klyukin

pgsql-hackers by date

Next:	From: Jim Nasby	Date: 2011-07-17 20:31:45
	Subject: Re: proposal: new contrib module plpgsql's embeded sql validator
Previous:	From: Andrew Dunstan	Date: 2011-07-17 20:25:53
	Subject: PostgreSQL Buildfarm client release 4.6