Calculate measure in DAX sum table A * count table B

[mizaskun]

Good morning,

I have the following simplified universe:

The OPINION table stores the opinion (from 1 to 10) of certain users for different questions (each user can answer just once):

The Countries table stores the countries in which each user is relevant:

I want a measure that calculates the average opinion of each user * count of the user in the table countries:

For example, if I had a table by question, the result shoud be calculated like this:

P1: User A (3 * 1) + User B (7 * 2) + User C (5 * 3) + User D (2 * 0)  / (1+2+3+0) = 5,33

P2: User A(9 * 1) + User B (5 * 2)  / (1+2) = 6,33

Any suggestion? 

Thank you so much! 

[PaulOlding · Accepted solution]

Here's a measure that returns your expected results for the example dataset

Average Opinion =

VAR _OpinionsWithCount =

ADDCOLUMNS(

OPINION,

"@Count", COUNTROWS(RELATEDTABLE(COUNTRIES))

)

VAR _Numerator = SUMX(_OpinionsWithCount, OPINION[VALUE] * [@Count])

VAR _Denominator = SUMX(_OpinionsWithCount, [@Count])

VAR _Result = DIVIDE(_Numerator, _Denominator)

RETURN

_Result

[Anonymous]

Your model is worse than you might think. You have to change it.

 <-- This is what should be visible in the Fields pane

And measures:

# User Countries = 
CALCULATE(
    DISTINCTCOUNT( UserCountryBridge[CountryID] ),
    REMOVEFILTERS( Countries )
)

Opinion Value = 
var UserAvgValueWithCountryCounts =
    ADDCOLUMNS(
        DISTINCT( Users[UserID] ),
        "@AverageValue",
            CALCULATE(
                AVERAGE( Opinions[Value] )
            ),
        "@CountryCount",
            CALCULATE(
                DISTINCTCOUNT( UserCountryBridge[CountryID] ),
                REMOVEFILTERS( Countries )
            )
    )
var Result = 
    DIVIDE(
        SUMX(
            UserAvgValueWithCountryCounts,
            [@AverageValue] * [@CountryCount]
        ),
        SUMX(
            UserAvgValueWithCountryCounts,
            [@CountryCount] * (1 - ISBLANK( [@AverageValue] ) )
        )
    )
return
    Result

Please do understand the logic of the measure in all the situations you'll be using it in. If in some of them you don't get what you need, then modify the measure to your heart's content.