From: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Rob Richardson <Rob(dot)Richardson(at)rad-con(dot)com>, pgsql-general(at)postgresql(dot)org |
Subject: | Re: Daylight savings time confusion |
Date: | 2010-03-16 15:11:59 |
Message-ID: | 20100316151159.GD3037@alvh.no-ip.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Tom Lane wrote:
> If my guesses are correct, then the minimum change to avoid this type
> of problem in the future is to change UTCTimestamp to be declared as
> timestamp WITHOUT time zone, so that you don't get two extra zone
> rotations in there. However, I would strongly suggest that you rethink
> how you're storing the data altogether. Two columns that represent the
> identical item of information is not good database design according to
> any theory I've ever heard. What I'd store is a single fire_date column
> that is of type timestamp with time zone and is just assigned directly
> from current_timestamp without any funny business. Internally it is UTC
> and completely unambiguous. Subsequently you can read it out in any
> time zone you want, either by setting TimeZone appropriately or by using
> the AT TIME ZONE construct to do a one-time conversion.
And possibly store the original timezone as a separate column, if that
information is of any value.
--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.
From | Date | Subject | |
---|---|---|---|
Next Message | Manlio Perillo | 2010-03-16 15:14:43 | Re: restoring a database to its initial state |
Previous Message | Arnaud Lesauvage | 2010-03-16 15:11:05 | Re: UPDATE with JOIN not using index |