Re: [HACKERS] Interval aggregate regression failure

Michael Glaesemann wrote:
>
> On Sep 4, 2006, at 4:45 , Bruce Momjian wrote:
>
> > Another question.  Is this result correct?
> >
> >     test=> select '999 months 999 days'::interval / 100;
> >             ?column?
> >     -------------------------
> >      9 mons 38 days 40:33:36
> >     (1 row)
> >
> > Should that be:
> >
> >      9 mons 39 days 16:33:36
>
> Yeah, I think it should be. I had been thinking of treating the month
> and day component as having separate time contributions, but it makes
> more sense to think of month as a collection of days first, integral
> or no, and then cascade down the fractional portion of the combined
> days component to time. I.e., 9.99 mon is 9 mon 29.7 days, rather
> than 9 mon 29 days 60480 sec.

No, I don't think so.  If we do that, there is no reason to cascade at
all.  Why not just say 9.1 months?

I am going to work on a patch to fix the >24 hours case, which will fix
your 24:00:00 case at the same time.  Will post in an hour.

--
  Bruce Momjian   bruce@momjian.us
  EnterpriseDB    http://www.enterprisedb.com

  + If your life is a hard drive, Christ can be your backup. +