On Sun, Jun 13, 2010 at 17:58, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Magnus Hagander <magnus@hagander.net> writes:
>> On Sun, Jun 13, 2010 at 17:42, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>>> ... (We presumably want
>>> timezone to default to the system timezone setting, but I wonder how
>>> we should make that work --- should an empty string be treated as
>>> meaning that?)
>
>> Umm. NULL could be made to mean that, or we could provicde two
>> different versions - one that takes TZ and one that doesn't.
>
> Using NULL like that seems a bit awkward: for one thing it'd mean the
> function couldn't be STRICT, and also it'd be bizarre that only this
> one argument could be null without leading to a null result.
Hmm, yeah.
> And two separate functions isn't good either. Basically, I think it's
> important that there be a way to specify an explicit parameter value
> that behaves identically to the default.
In that case, empty string seems fairly reasonable - if you look at
the text based parsing, that's what we do if the timezone is an "empty
string" (meaning not specified).
>> And a third, construct_time(), no?
>
> Yeah, maybe ... do you think there's any demand for it?
Yes, I think there is. Plus, it's for completeness :-)
--
Magnus Hagander
Me: http://www.hagander.net/
Work: http://www.redpill-linpro.com/