I'm sure this is possible, I'm just struggling to find a way to do it!
I've got a transaction volume cube that is dimensioned by day & customer which counts the number of transactions per day.
The data feeding this is:
Transaction_ID Customer_ID Date Value
1000000001 8000000001 20180101 1
1000000002 8000000002 20180101 1
1000000003 8000000001 20180101 1
I can do a count of transactions by day, week, month, year etc but what I am try to do is count the number of customers that have made transactions by day, but also have this where date filters can work so if I want to put month on the axis instead of day, it still does a distinct count.
So for the example above, 2 customers made 3 transactions on 1st January 2018.
Is there any way of doing this?