Re: strange TIME behaviour

Martijn van Oosterhout <kleptog(at)svana(dot)org> writes:
> Historical I beleive. Postgres has four types: timestamp, timestamptz,
> time and timetz. Then SQL decreed that TIMESTAMP means WITH TIME ZONE,
> ie timestamptz. So now you get the odd situation where:

> timestamp == timestamp with time zone == timestamptz
> "timestamp" == timestamp without time zone == timestamp
> time == time without timezone

This isn't correct --- timestamp has meant timestamp without time zone
for a long time (since 7.3 I believe). Once upon a time it worked like
you show here, but we changed it specifically because the SQL spec says
that WITHOUT TIME ZONE is the default.

In the case of TIME, that's a good default; in the case of TIMESTAMP
not so much, but we're stuck with it because the spec says so.

regards, tom lane