Hierarchies of distincts

Question

Dear all,
I have to create a cube where I count my active customers per timeframe. The simplest design would be to put a line for each active customer linked to a time dimension into the cube.
Something like
customer_no(dimension_key for customer dimension) active_date(dimension_key for time dimension)
In my reporting tool I could then create the different distinct calculations for the different timeframes.
But I wonder if I could also use the hierarchy mechanism of OWB. I would load the cube with the daily distincts and the month with the monthly distincts and so on.
I just have to ensure, that a customer that is active on two days in a month is not counted twice in the distinct customers per month calcualtion.
Any ideas?
Regards,
Knut

Answer 1

Hi Scooby,
thanks for your reply. I am doing exactly what you say, but this doesn't work for distinct counts. Let me try an example:
customer_no active_date
1 1/1/10
2 1/1/10
3 1/1/10
1 1/2/10
1 1/2/10
3 1/2/10
result for daily distinct customers
1/1/10 3
1/2/10 2
result for monthly distinct customers
Jan 2010 5
and here the problem occurs. There were 3 distinct customers in January, not 5.
Regards,
Knut
Edited by: knherzog on 13.07.2010 04:04