Unanswered: Advise about design of a measure in OLAP
I need your advise about a performance problem that i'm having with a calculated member in a Cube.
I did a calculated member that show a Avg of set of values on throught time(Yearly, Monthly and Weekly), it did its calculates using another calculated member(this other calculated member calculated show AVG too pn thought rotation works hours ("Turnos") -well, AvG on AVG is a crap but that is the bussiness logic of my customer-).
My problem es when I try to visualize the data in teh cube this consume a lot of time in to show its values (the cause is the first calculated member taht I explain up). I think in some possibles solutions:
1. Put this calculated member as a measure in my fact table, but the problem is taht it is a Avg and then I would have that calculate all possible combianation of my dimension for the Avg. I'm usin 8 dimension.
2. I try to define my cube as ROLAP, my problem still persit, my measure is slow.
3. I try to define a derived measure but I can't 'cause if I use Avg(<colum of my fact table>) how my source colum fails the process of calc (the others columns in my fact no exist in a clause group by).
I'm open to read any suggestion about that. I know that is not enough information but star with that and if someone has any question I could put more data about my case.
Thxs in advance