How the Filter definition of the metric can affect the final result.

[db2907]

Let's assume that I have the Products table in my model defined like this:

ProductID	UnitPrice	UnitsInStock	Discontinued
1	10	50	TRUE
2	15	30	FALSE
3	20	20	TRUE
4	25	10	TRUE
5	5	15	FALSE

I have created two metrics :

Value1  = 

CALCULATE ( SUMX ( Products, Products[UnitPrice] * Products[UnitsInStock] ), FILTER ( Products, Products[Discontinued] = TRUE ) )

Value2 = 

CALCULATE ( SUMX ( Products, Products[UnitPrice] * Products[UnitsInStock] ), Products[Discontinued] = TRUE )

The only difference is the definition of the filter. I have evaluated the results in Power Desktop but I do not see any difference. This leads to my question: How could this definition affect the final result of the metric or when might this occur?

[johnt75 · Accepted solution]

The two measures are essentially identical, both are calculating the result for products which are discontinued. 

The definition for value2 is rewritten internally to be

Value2 =
CALCULATE (
    SUMX ( Products, Products[UnitPrice] * Products[UnitsInStock] ),
    FILTER ( ALL ( Products[Discontinued] ), Products[Discontinued] = TRUE )
)

The form of CALCULATE where you can specify a predicate as a filter rather than a table is merely syntax sugar, to make the code easier to read and write.