Re: Yet Another Timestamp Question: Time Defaults

On Monday, January 21, 2013 at 18:11, bgd39h5xxt@sneakemail.com
(Nathan Clayton nathanclayton-at-gmail.com |pg-gts/Basic|) wrote:

>I only wish. I work with a transactional system from the 70s on
>a daily basis that decided to store something like a "work date" and
>"work time". The date changes whenever they decide to dateroll the
>system. Until then the time field continues to grow, so you see times
>like 25:00 and 26:00 all the time.
>
SELECT execute(relevant_dba) FROM the_70s WITH tardis WHERE
working_tardis = true;

>Exceptions abound.
>
At least that can't be blamed on a government, and, we can only
hope ISO-8601 will prevent more examples being created.

You sound as though you really need, and/or already have, a
dedicated datatype... if only to stop 'the system' from 'fixing'
such weirdness.

Regards
Gavan Schneider