Re: Interval Precision

Steve Crawford <scrawford(at)pinpointresearch(dot)com> writes:
> Check the archive of the HACKERS list for messages with the subject
> "Bug 1500".

> I don't see that any absolute decision has been reached but the
> majority opinion seems to be to fix the problems with interval
> formatting and possibly also extend/improve it rather than to kill it
> off.

I don't think there's any way that we'll remove it until a better
implementation is available. Karel was essentially arguing "if you
take it away then someone will be motivated to write the better
version", but we don't generally work that way ...

The basic issue is that it's sharing code with to_char(timestamp)
and therefore accepts a whole bunch of format codes that don't really
make sense for intervals; and at the same time is lacking some that
do make sense. So when it does get fixed you might have some issues
with format codes going away ... but the ones that are actually useful
with intervals, like HH and MI, presumably will still be there.

regards, tom lane