On Thu, Jun 01, 2006 at 12:35:44PM -0400, Tom Lane wrote:
> Joachim Wieland <joe@mcknight.de> writes:
> > I'm talking about the timetz type that does not carry a date. So you don't
> > know if daylight savings time is active or not. How would you interpret the
> > full timezone in this case without a date?
> Oh, doh, I managed to miss that detail. Yeah, you're right, you need an
> arbitrary assumption in that case. Or we could forbid these timezones
> in timetz input, but that's probably not very helpful.
After sending my last mail, I concluded that it was in fact me who missed
something and that you were right. I came to the conclusion that you were
talking about the fact that you can specify a timetz also with a date:
template1=# select '2006-06-01 10:49 America/New_York'::timetz;
timetz
-------------
10:49:00-04
This date can then be used to infer the timezone:
template1=# select '2006-03-01 10:49 America/New_York'::timetz;
timetz
-------------
10:49:00-05
I have updated my patch to do so. Just specifying a timestamp
select '10:49 America/New_York'::timetz;
does now return an error.
Is that a suitable compromise?
Joachim