Forum Discussion

RJ's avatar
RJ
Resolver II
7 years ago
Solved

DAX using IF (aggregation within another aggregation). I get an incorrect total.

  What I want is the total to equal 2,947,891 for Sumif and Measure 2   The three measures were   Sales£OrdDate2   Sales£OrdDate2 = CALCULATE([Sales£M] , USERELATIONSHIP(Header[Order Da...
Need Help
  • RJ's avatar
    RJ
    7 years ago

    Brilliant

     

    It worked. Thank you

     

    Sumif4 =
    VAR tableA = summarize(Dates,Dates[MonthYear],"a",
    CALCULATE([Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] )) )
     
    RETURN
    CALCULATE ( [Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] ) , FILTER ( tableA, [a] > 100000 ) )
     
    //-----------------------
     
    So a variable is set up called VAR tableA
    This generates a temp table summarized by a column Dates[MonthYear]. With the measure 
    CALCULATE([Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] ))
     
    So in effect a  temporary table is created. This table is called TableA. It can be called any name. This temp table has one dimension >>Dates,Dates[MonthYear]  << and one measure called a. It could however be called any name required
     
    This temp table is then used to filter the expression 
    CALCULATE ( [Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] )
    as required
     
    NB. Another option is to create a new actual table. Say
     
    TableName =
    SalesMthYrOrdDate = ADDCOLUMNS( summarize(Dates, Dates[MonthYear],Dates[Year]),   //load required dimensions
    "SalesOrdDate",                                //or whatever name you wish to call the measure
    CALCULATE([Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] )) )   //measure column called "SalesOrdDate"
     
    and then join the MonthYear column to the MonthYear column in the Dates table //link using PBI relationships
     
    Final step. Use this expression
     
    Sumif7 =
    CALCULATE ( [Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] ) , FILTER (SalesMthYrOrdDate,SalesMthYrOrdDate[SalesOrdDate] > 100000 ))
     
  • RJ's avatar
    RJ
    7 years ago

    To make it more effecient. Add 'ADDCOLUMN'

     

    Sumif6 =
    VAR tableB = ADDCOLUMNS( summarize(Dates, Dates[MonthYear],Dates[Year]), "SumMthYr",CALCULATE([Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] )) )
    RETURN CALCULATE ( [Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] ) , FILTER ( tableB, [SumMthYr] > 100000 ) )
     
     
     
    or an alternative is
     
    Sumif6 =
    VAR tableB = summarizecolumns(Dates, Dates[MonthYear],Dates[Year], "SumMthYr",CALCULATE([Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] )) )
    RETURN CALCULATE ( [Sales£M] ,USERELATIONSHIP(Header[Order Date] ,Dates[Date] ) , FILTER ( tableB, [SumMthYr] > 100000 ) )
Ashish_Mathur's avatar
Ashish_Mathur
Super User
7 years ago

Hi,

 

Does this work?

 

=CALCULATE(SUMX(VALUES(Calendar[Month-Year]),[Sales£M]),FILTER(Header,[Sales£M]>0),USERELATIONSHIP(Header[Order Date] , Dates[Date]))

RJ's avatar
RJ
Resolver II
7 years ago

Thanks Ashish

 

I will try this when I return.

 

Its an aggregation within an aggregation issue but Power BI must have an solution for this. But I will try your measure later