Re: [office-formula] DATEDIF (Section 6.9.2)

["Andreas J. Guelzow" <aguelzow@math.concordia.ab.ca>]

On Wed, 2009-02-25 at 14:53 +0100, Eike Rathke wrote:
> Hi Andreas,
> 
> On Friday, 2009-02-13 12:46:57 -0700, Andreas J. Guelzow wrote:
> 
> > I think the real problem here is that the so called "definition" doesn't
> > really define the behaviour:
> > 
> > =DATEDIF(DATE(1992;2;15); DATE(1993;9;15); "d")
> > 
> > Are we counting days here in 1992 (a leap year) or in 1993 (not a leap
> > year)?

Hmm, yeah I meant "yd". (For some reason my mail messages frequently
don't contain what I mean. I hope computers will improve some day!)
> 
> "d" is said to calculate the difference in days. From my understanding,
> the real day count should be calculated, so the answer should be equal
> to
> 
> =DATE(1993;9;15)-DATE(1992;2;15)    => 578

That's not the number of days between those two days. It is the
difference between the date serial numbers.

There is no day between DATE(1993;9;15) and DATE(1993;9;16), but the
difference is 1.

(I think we should use the difference, and not say "between".)
> 
> which is what Gnumeric does also for DATEDIF().
> 
> Assuming you wanted refer
> 
> =DATEDIF(DATE(1992;2;15); DATE(1993;9;15); "yd")

yes that's what I meant.

> 
> instead, that indeed lacks definition. Gnumeric's result is 212, so
> calculates days in 1993 (non-leap year). The rule behind that seems to
> be "take (non-)leap year of the second date", as
> 
> =DATEDIF(DATE(1991;2;15); DATE(1992;9;15); "yd")
> 
> results in 213 instead. Can someone confirm / deny?

again even if we use the second (or perhaps later year) we should not
talk about "between".

Andreas
-- 
Andreas J. Guelzow, PhD, FTICA
Coordinator, Mathematical & Computing Sciences
Concordia University College of Alberta