Default selections overruling set analysis exclusion

[Spiller]

Hi all,

I had to copy below KPI numerator to an additional dashboard to use it there as well. Only at the new dashboard we have a default bookmark with calendar year set at 2022 and ProgramActive at No. I would have expected that the these selections would have been disregarded in the set analysis. But i was wrong. Can anyone tell what i am overlooking? Doesn't it work with the aggr + if combination that i am using now? I can't remove the default bookmark selection from the dashboard itself as it's for other data.

=Sum(Aggr(If(Max({<

 TaskName = {'Task1'}
,ProgramActive
,CalendarYear
,CalendarQuarterNumber
,CalendarMonthLabelShort
,TaskStatus = {'Completed'}
,ProjectLeading = {'Yes'}
>} Delta ) > Threshold
,1
), ProjectCode)
)

Thanks in advance.

Spiller

[marcus_sommer]

Try it with the condition in the outer aggregation, too. Something like:

sum({ < ... >} aggr(if(max({< ... >} value)> x, 1), Dim))

- Marcus

[Spiller]

@marcus_sommer Thanks, tried that as well, but still the same effect. For the denominator part, which is a count distinct of the projects and same dimensions i have no issues.

[marcus_sommer · Accepted solution]

Maybe you need to include further fields within the set analysis which are included from the bookmark, too. In this regard you should also look for invisible selections which may occur if the fields are tagged with 'hidden' or defined with hideprefix/hidesuffix.

Another look may go for alternate states. If there any and/or they are different between both application it may change which selections are considered.

Further you compares your results against Threshold which may also react on the selections.

- Marcus 

[Spiller]

Thanks for that last line, as the Threshold was indeed the root cause. It's based on the business. Should I make a variable out of that one then?

[marcus_sommer]

It depends. If Threshold would be as variable a fixed value it should work but if it's a calculation it wouldn't change the behaviour unless you could there also apply a set analysis which ignored the unwanted selections.

- Marcus

[Spiller]

thanks, i have sorted it now