Do not allow a measure element to be filtered

[StanleyBlack]

Afternoon everyone,

I have a load of tiered measures and have run into a problem. I use this to calculate the number of students below 90%

Persistent Absence = 
VAR __ISINSCOPE = ISINSCOPE ( Attendance[Year] )
VAR __PACore = CALCULATE(DISTINCTCOUNT(Attendance[UPN-SUID]),FILTER(
                VALUES('Attendance'[UPN-SUID]),
                            [Attendance Percentage YTD] <= 0.90))

RETURN

IF (
    __ISINSCOPE,
        __PACore,
        CALCULATE(__PACore,NOT('Attendance'[Year] IN {"Nursery", "Reception"}                     
)))

I then use the following to find the percentage:

Persistent Absence % = Divide([Persistent Absence YTD], DISTINCTCOUNT(Attendance[UPN-SUID]))

I then have multiple measures that give me breakdowns of the percentage such as:

Male Persistent Absence % YTD = 
Calculate([Persistent Absence % YTD], 'Student Database (New)'[Gender] IN {"Male"})

The problem is that I only want to filter the numerator of the second measure by the 3rd and not the denominator. Is there a way to make:

Persistent Absence % = Divide([Persistent Absence YTD], DISTINCTCOUNT(Attendance[UPN-SUID]))

the element in bold not filterable by the 'Student Database (New)' table?

The alternative is editing multiple measures rather than just one.

Thanks,

Stan

[TomMartens]

Hey,

not if I understand your question correctly, but why don't you just rewrite this

Persistent Absence % = Divide([Persistent Absence YTD], DISTINCTCOUNT(Attendance[UPN-SUID]))

to this

Persistent Absence % = Divide([Persistent Absence YTD], CALCULATE(DISTINCTCOUNT(Attendance[UPN-SUID]) , ALL('Student Database (New)')))

Regards,
Tom

[StanleyBlack]

Hi Tom,

Sorry if that was less than clear.

Say I have 1000 kids in a school and 100 are below 90%.

The second calculation would be 100/1000 and = 10%

When I add a filter on the third measure it changes everything so for males say I had 50 below 90% and had 500 males in the school the third measure currently returns:

50/500 when it should return 50/1000 which is the total school population rather than the subgroup total.

Using the measure you suggested seems to make everything way off.