Forum Discussion
sum or count a measure
- 8 years ago
Thanks for that.
It looks like the problem is the VAR/RETURN syntax.
If you use VAR to store the result of your existing measure like this, it actually just evaluates once and locks in a single value, evaluated in the overall filter context (i.e. not per week).
We need the validobservation measure to be re-evaluated for every iteration of the SUMX, so the SUMX should refer to a measure (or equivalent expression wrapped in CALCULATE), not a variable.
I would say the cleanest way to do this is to keep your first measure as you have defined it (since you know it's working at the Week level), and call it say validobservation.
Then your new measure should be:
Valid Week Count = SUMX ( VALUES ( TimeTable[Week] ), [validobservation] )That is, the second argument of the SUMX should be a measure, not a variable.
Does that help?
Regards,
Owen
Thanks for that.
It looks like the problem is the VAR/RETURN syntax.
If you use VAR to store the result of your existing measure like this, it actually just evaluates once and locks in a single value, evaluated in the overall filter context (i.e. not per week).
We need the validobservation measure to be re-evaluated for every iteration of the SUMX, so the SUMX should refer to a measure (or equivalent expression wrapped in CALCULATE), not a variable.
I would say the cleanest way to do this is to keep your first measure as you have defined it (since you know it's working at the Week level), and call it say validobservation.
Then your new measure should be:
Valid Week Count =
SUMX (
VALUES ( TimeTable[Week] ),
[validobservation]
)That is, the second argument of the SUMX should be a measure, not a variable.
Does that help?
Regards,
Owen
yes it works well now!
thank you very much