Hi folks,
Thanks for your feedback, guys. Tom, you're right, I would not have used timetz, except for the fact, that the current_time returns exactly that. But I guess, that's not a driver problem. So, Oliver, I understand your argument, it sounds logical. I guess then, it's a feature, not a bug ;-)
Cheers
Lukas
2010/9/19 Tom Lane
<tgl@sss.pgh.pa.us>Oliver Jowett <
oliver@opencloud.com> writes:
> Did you see Kris's earlier response here? See
>
http://archives.postgresql.org/pgsql-jdbc/2010-05/msg00052.php. The
> problem is we need to pass around a timezone offset, but JDBC +
> java.util.Date give us no way to do that without subclassing those types
> (which seems a bit hairy). Without that extra data, timetz just doesn't
> map well to any of the standard Java date/time types.
timetz is a fundamentally brain-dead data type to start with ---
it simply doesn't carry enough information to deal with timezones
meaningfully, at least not once you start considering DST changes.
This is the SQL standard's fault not ours, so there's not a lot
we can do about it other than recommend people avoid timetz.
regards, tom lane