1 of 1 people found this helpful
Hi Bob -
In order to get a population counts that aren't impacted by anything on your filter shelf, try using FIXED instead of EXCLUDE in your denominator (check out the order of operations if you haven't already).
When you use FIXED, you will want to specify a dimension before the colon, otherwise the count will be across your entire dataset.
If you want to use multiple demographic attributes, instead of using Gender, you could create a parameterized dimension (the link is for measures, but same process for dimensions).
I don't know if you can open attached. It is in 10.3, But here is the Solution:
Pl Check Screenshot and attached.
Here are LODs used:
cohort_test.twbx 15.4 KB
Using FIXED [Gender] makes it look right when I initially add [Date] to the Filters shelf, but actually changing the date range is not having the intended effect on the numerator. For example, with the full date range (6/1/2017 - 9/1/2017), I correctly see:
Female - 1/2
Male - 2/3
But if I change the range to 6/1/2017 - 7/1/2017, the Male bar stays at 2/3 instead of changing to 1/3.
And if I change it to 6/1/2017 - 6/1/2017, the Female bar disappears completely. In that case, I want to see:
Female - 0/2
Male - 1/3