[ http://tools.oasis-open.org/issues/browse/OFFICE-2630?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18371#action_18371 ]
Dennis Hamilton commented on OFFICE-2630:
-----------------------------------------
Considering the number of other financial functions that use YEARFRAC internally (one would hope) to deal with their basis adjustments, we have a definite challenge in front of us.
This is looking more and more like a great Codefest challenge problem.
It appears that OOo 3.0 is computing the result as 1+29/365 with no accounting for the fact that 2008 is a leap year. Surprise, surprise. My TI BA-35 calculator matches them to that formula exactly out to 8 decimal places with correct rounding in the 9th.
The Gnumeric and Excel 2007 result is (334+60)/365.5 = 1.077975376 (by calculator), operating on the assumption, apparently, that the spanned-year durations are averaged. (I have seen that in basis descriptions, I think, though I never explored the consequences very closely. I think David Wheeler's extensive tests confirmed something like this.)
I wonder if there are discontinuities in the actual/actual basis rule, where adding a day in a new (leap) year causes the YearFrac to be less than the period ending one day earlier (or starting one day later)..
> YEARFRAC ambiguity
> ------------------
>
> Key: OFFICE-2630
> URL: http://tools.oasis-open.org/issues/browse/OFFICE-2630
> Project: OASIS Open Document Format for Office Applications (OpenDocument) TC
> Issue Type: Bug
> Components: OpenFormula
> Affects Versions: ODF 1.2 Part 2 CD 1
> Reporter: Robert Weir
>
> I thought we nailed this one a while ago, but I see the problem still in Part 2, CD 01.
> YEARFRAC is currently defined as "Computes the fraction of the number of years between a StartDate and EndDate.
> Basis is the system for determining how many days are in a month or year."
> OK. So YEARFRAC is A/B where A= number of days between the two dates, using the counting conventions of the basis. And B is what? A year certainly. But how many days? 365? 365.25? Dependent on whether date 1 or date 2 is in a leap year?
> I thought David figured this out at one point.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://tools.oasis-open.org/issues/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
