Hhhhmmm, I’m trying to puzzle something out. I’ve just got a cube that we’ve started, and it’s based on payperiod. Usually there’s 26 payperiods in a year of course, so that’s how I set up the dimension. But this year, because of how the calendar works, we’ve got, yep, you guessed it, 27 payperiods.
So I’m not quite sure how / what to do with that extra data. And I know for sure VPFIN is going to ask me at some point soon.
Michael, what would be the downside of adding the 27th pay period and ignoring it for most years?
If the problem is spreading, perhaps create an extra parent level for the 27th period, so that spreading on most years will affect the 26 periods only, and for strange years that have 27, you spread at the higher level.
It takes some management and administration to keep it straight, but that should be limited to once per year. And it seems like it would keep the integrity of comparing period-level statistics regardless of whether it’s 26- or 27-period year.
Annual statistics may be affected, but that’s the nature of the beast. You really are laying out more cash for payroll in a 27-period year, but you should be able to pull averages based on period counts, etc. (I’d likely add a member variable to the time dimension to record the number of periods for each year, just to make it convenient for calculations or reporting; but I have to confess I’m answering from my phone so I can’t check this but it occurs to me that member variables on a time dimension might not be available.)
Thank you for sharing this case Michael and thanks to all of you for providing workarounds and tips.
Michael could you please add this request (Including the case) to the feedback forum? This will be more visible for the product team to analyze the request.