If you have a Slowly Changing Dimension (SCD) Type 2 for your customer and you want to calculate the number of distinct customers that bought a product, you cannot use the simple formula:
Customers := DISTINCTCOUNT( FactTable[Customer Id] ) )
because it would return the number of distinct versions of customers. What you really want to do is to calculate the number of distinct application keys of the customers, that could be a lower number than the number you’ve got with the previous formula. Assuming that a Customer Code column in the Customers dimension contains the application key, you should use the following DAX formula:
Be careful: only the version above is really fast, because it is solved by xVelocity (formerly known as VertiPaq) engine. Other formulas involving nested calculations might be more complex and move computation to the formula engine, resulting in slower query.
This is absolutely an interesting pattern and I have to say it’s a killer feature. Try to do the same in Multidimensional…
Counts the number of distinct values in a column.
DISTINCTCOUNT ( <ColumnName> )
Counts the number of rows in a table.
COUNTROWS ( <Table> )
Creates a summary of the input table grouped by the specified columns.
SUMMARIZE ( <Table> [, <GroupBy_ColumnName> [, [<Name>] [, [<Expression>] [, <GroupBy_ColumnName> [, [<Name>] [, [<Expression>] [, … ] ] ] ] ] ] ] )