Sum elements in a table discriminating one dimension

[Numerozero]

Hello,

I am looking for a way to sum elements in a table discriminating one dimension while including another dimension.
Here in this example, I want the sum of errors for the "Jean" and "Shirt" calculated separately, while the months of Jan and Feb are compensating each other :

		Ventes	Forecast	error
Pant	jan	10	0	10
Pant	feb	0	15	15
Shirt	jan	20	0	20
Shirt	feb	10	10	0
		40	25	25

Here we can see the error is only 25 units, because there is only 5 units of error for "Pant" (Jan and Feb are compensating each other), while the error for "Shirt" is 20 units.

Is there any way in DAX to reach this result ?

Thank you

[Jihwan_Kim · Accepted solution]

Hi,

Please check the below picture and the attached pbix file.

Ventes measure: = 
CALCULATE( SUM(Data[Value]), Data[Attribute] = "Ventes")

Forecast measure: = 
CALCULATE( SUM(Data[Value]), Data[Attribute] = "Forecast")

Error measure: =
IF (
    ISINSCOPE ( 'Month'[Month] ),
    ABS ( [Ventes measure:] - [Forecast measure:] ),
    SUMX (
        DISTINCT ( Category[Category] ),
        CALCULATE (
            ABS ( [Ventes measure:] - [Forecast measure:] ),
            ALL ( 'Month'[Month No], 'Month'[Month] )
        )
    )
)

[Numerozero]

Adapted the code to my needs and it worked great. Thanks