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 Date] , Dates[Date]))
 
Sum if
 
Sum If = if    (CALCULATE([Sales£M] , USERELATIONSHIP(Header[Order Date] , Dates[Date]))
>100000
, CALCULATE([Sales£M] , USERELATIONSHIP(Header[Order Date] , Dates[Date])) )
 
Measure 2
 
Measure 2 = switch(true() , CALCULATE([Sales£M] , USERELATIONSHIP(Header[Order Date] , Dates[Date]))
> 100000,
CALCULATE([Sales£M] , USERELATIONSHIP(Header[Order Date] , Dates[Date])) )
 
Thanks for any help on this. Hopefully there is a simple solution
 
and 
 
Sales£M = sum (SalesDetail[Sales Value £])
 
 
 
 
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 ) )

8 Replies

  • RJ's avatar
    RJ
    Resolver II
    7 years ago

    Ill try this tomorrow

     

    https://community.powerbi.com/t5/Desktop/DAX-Measure-with-Nested-IF-Statements/m-p/113358/highlight/true#M47827

     

    DAX Measure with Nested IF Statements#

     

    Re: DAX Measure with Nested IF Statements

    [ New ]
     
    Options
     

    ‎01-14-2017 08:04 PM - edited ‎01-14-2017 08:08 PM

    @MWinter225If you do want Measures - these should work also! Smiley Happy

    MEASURE 1

    Total Adj Sales ALT =
SUMX (
    'Table',
    IF (
        'Table'[Adjustment] = "b",
        'Table'[Sales] * 0.9,
        IF ( 'Table'[Adjustment] = "c", 'Table'[sales] * 0.5, 'Table'[Sales] )
    )
)

    MEASURE 2 - SWITCH is internally converted into nested IFs - one thing I really like is that its much easier to read and write

    Total Adj Sales ALT 2 =
SUMX (
    'Table',
    SWITCH (
        TRUE (),
        'Table'[Adjustment] = "b", 'Table'[Sales] * 0.9,
        'Table'[Adjustment] = "c", 'Table'[sales] * 0.5,
        'Table'[Sales]
    )
)

    Now you have 3 options which should all give you the same result!

     

    Good Luck! Smiley Happy

    [

  • 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