How to calculate Cumulative Total where the value to be Totaled is a measure (rather than a column)

[stfox]

Hi Folks

I am having issues calculating a Cumulative Total variance - the cumulative difference between Sales and Targets

The standard calculation for a cumulative total is outlined below.  

Cumulative_Total =
    CALCULATE (
        SUM ( Table[Column1] ),
        FILTER ( ALL ( Table[Day] ), Table[Day] <= MAX ( ( Table[Day] ) ) )
    )

But this method only accepts a column as the value to be summed (in bold above) . If a measure is used  the following error is produced :

The SUM function only accepts a column reference as the arguement number 1.

I have a situation where I want to calculate Cumulative Total variance between Sales and Targets.  Sales and Targets sit in different tables. Sales are recorded at a more granular level than Targets.  Is there a way to created a Cumulative total variance in this situation? 

Cheers

Steve 

[Framet]

Hi,

Would is not be the case that the cumulative variance of each row would be the same as the variance between the total sum of sales and sum of targets. Can you do each calculation and simply subtract one from another?

Cumulative_Total =
    CALCULATE (
        SUM ( SalesTable[Sales] ),
        FILTER ( ALL ( Table[Day] ), Table[Day] <= MAX ( ( Table[Day] ) ) )
    )
        - CALCULATE (
            SUM ( TargetsTable[Targets] ),
            FILTER ( ALL ( Table[Day] ), Table[Day] <= MAX ( ( Table[Day] ) ) )
        )

I could be completly missing the point and your data prevents this from making sense. Could you provide a little more on how the tables relate and their structure?

You might find SUMX works for you allowing you to sum a performed a calculation for every row context of the indicated table?

Thanks

Thomas

[stfox]

Thanks Thomas,

 Subtracting the Cumulative Sales from the Cumulative Targets  totally makes sence. It was my first port of call (I should have mentioned it).  But strangely it doesn't generate the correct result (screenshot below and actual file attached in this link ). I am thinking I am missing something simple . 

I don't understand the changing from SUM to SUMX logic (incidentally I also get an error )- can you elaborate?  

Much appreciated.

Steve

[Anonymous · Accepted solution]

Hi @stfox,

From your PBIX file, I note that when creating Cumulative Sales, you use the following formula.

Cumulative Sales = CALCULATE (
    SUM ('Sales'[#Sold] ),
    FILTER (
        ALL ( 'DailyTarget2'[Date]),
        'DailyTarget2'[Date] <= MAX ( 'DailyTarget2'[Date])
    )
)

But when creating Test_Cumulative_Variance_Wrong, you use a different formula to represent Cumulative Sales.

CALCULATE (
    SUM ('Sales'[#Sold] ),
    FILTER (
        ALL ( 'DailyTarget2'[DailySalesTarget] ),
        'DailyTarget2'[DailySalesTarget] <= MAX ( 'DailyTarget2'[DailySalesTarget])
    )
)

When you change the formula of Test_Cumulative_Variance_Wrong as follow, does it give your expected result?

Test_Cumulative_Variance_Wrong = (CALCULATE (
    SUM ('Sales'[#Sold] ),
     FILTER (
        ALL ( 'DailyTarget2'[Date]),
        'DailyTarget2'[Date] <= MAX ( 'DailyTarget2'[Date])
    )
))-
 (CALCULATE (
    SUM ('DailyTarget2'[DailySalesTarget]),
    FILTER (
        ALL ( 'DailyTarget2'[Date] ),
        'DailyTarget2'[Date] <= MAX ( 'DailyTarget2'[Date])
    )
))

Thanks,
Lydia Zhang