Measure (LY variation)

Hi setis 

If I remember your data model correctly, you'll probably need something like the following. Do bear in mind that for time intelligence functions to work properly, the date table should always have complete years and contain all the dates that you need. I.e. if you have a date table from 01/01/2017 on,  SAMEPERIODLASTYEAR(anything in 2017) will return blank since the date table has no dates in 2016.

In any case, why not just use the date table to select the year you want?

NrCases Payment&NOFee =
CALCULATE (
    COUNTROWS (
        EXCEPT (
            FILTER (
                DISTINCT ( Financial_Combined[Job No.] );
                CALCULATE ( SUM ( Financial_Combined[Amount Paid] ) ) > 0
            );
            FILTER ( DISTINCT ( Fees[CaseID] ); CALCULATE ( SUM ( Fees[FeeAmount] ) ) > 0 )
        )
    );
    SAMEPERIODLASTYEAR ( 'Calendar'[Date] )
)

setis 

You can always create a second measure based on the first one for the last year, i.e., you have the initial measure:

NrCases Payment&NOFee =
COUNTROWS (
    EXCEPT (
        FILTER (
            DISTINCT ( Financial_Combined[Job No.] );
            CALCULATE ( SUM ( Financial_Combined[Amount Paid] ) ) > 0
        );
        FILTER ( DISTINCT ( Fees[CaseID] ); CALCULATE ( SUM ( Fees[FeeAmount] ) ) > 0 )
    )
)

and then, based on that:

NrCases Payment&NOFee LY =
CALCULATE ( [NrCases Payment&NOFee]; SAMEPERIODLASTYEAR ( 'Calendar'[Date] ) )

It's probably neater.

Cheers