I need a measures to give distinct counts over various dimensions. The column(lets call it col_id) on which the counts are based on is not required as a dimension/level. But I embed it as a level only for using category counts and later suppress that complete category. Also, because there is no direct Activity Measure associated with this, I create a dummy column in IMR/IQD to hold constant value '1'.
Now the problem is that this col_id field has over 200k values and it grows rapidly. So the cube generation is slow and also violates the category ratio rule. Is there any efficient way of creating this distinct count.
(I went through the archives and there were suggestions of doing this is through Impromptu; if thats the answer, please explain how?)