cancel
Showing results for 
Search instead for 
Did you mean: 
Reply
Highlighted
luiscb Regular Visitor
Regular Visitor

Minimum value problem with null value on high level aggregation

 

Hello everyone,

I have the following code

Conditions_Legend = 
var X = SUMMARIZE(fact,fact[country],fact[WeekText],"MIN", RANDBETWEEN(1,40)+0)
var Min_Country = MINX(X,[MIN])

return
Min_Country


Which applied to a matrix with WeekText as row headers and Country as Column Header produces the following:

 

table_countries.PNG

 

For Japan is NULL because for that specific week it does not hold data and my objective is when I take the country dimension from the matrix header, that the minimum retrieved to be 0 due to Japan.

Unfortunately, when I take the country dimension of the Matrix, it does not recognize the null value as 0 and so puts the minimum number value, which is 5

countries_bad.PNG

I have played around with if statement to convert null to zero, but I end up having the same result.
Any ideas on how not to ignore the null value in the condition above?

Thanks!
Ps: I used random values and added a 0 to try and force a 0 to appear instead of null

 

1 ACCEPTED SOLUTION

Accepted Solutions
luiscb Regular Visitor
Regular Visitor

Re: Minimum value problem with null value on high level aggregation

I have actually used the following approach and did what the trick:

var = SUMX(fact,SWITCH(TRUE(),ISBLANK([KPI Measure]),1,0))

So this problem is closed
2 REPLIES 2
Cmcmahan New Contributor
New Contributor

Re: Minimum value problem with null value on high level aggregation

I would use IF(ISBLANK()) here.

 

Something along the lines of 

var Min_Country = IF(ISBLANK(MINX(X,[MIN])),0,MINX(X,[MIN]))

If it's blank, you give it a default value of zero, otherwise return the normal value.

luiscb Regular Visitor
Regular Visitor

Re: Minimum value problem with null value on high level aggregation

I have actually used the following approach and did what the trick:

var = SUMX(fact,SWITCH(TRUE(),ISBLANK([KPI Measure]),1,0))

So this problem is closed