Forum Discussion
Measure for counting ignoring context filters
Hello,
I want to (distinctly) count how many stores have participated in the delivery of an order. I want the measure to ignore the specific filtercontext for the calculations only and not for the visual itself. The model is a Direct Model on an Analysis Service using a Star Scheme. An example of the data:
In the above picture i want it to write 3 on every row whilst only showing the 3 stores in question. Using AllExcept, ALLSELECTED etc. either does the same as the picture or counts every single "Store No." in the dataset. What i want is to tag every "Order No." with the number of unique "Store No.".
In SQL i would do it like in the picture below, but i do not know how in DAX.
Any help would be appreciated!
Hi JacobCoops ,
According to your description, Order No and Store No are in different tables, I create a sample, they have relationship by the Key column.
Here's my solution, create a measure.
Measure = CALCULATE ( DISTINCTCOUNT ( 'Store'[Store No] ), FILTER ( ALL ( 'Order' ), 'Order'[Order No] = MAX ( 'Order'[Order No] ) ), FILTER ( ALL ( 'Store' ), 'Store'[Key] = MAX ( 'Store'[Key] ) ) )Get the expected result.
I attach my sample below for reference.
Best Regards,
Community Support Team _ kalyjIf this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
11 Replies
- smpa01
Community Champion
JacobCoops the equivalent dax is following
Measure = CALCULATE(COUNT('Table'[storeNumber]),ALLEXCEPT('Table','Table'[orderNumber]))- JacobCoopsFrequent Visitor
Hi,
That return the following (there can be multiple rows per "Order No." and "Store No." combination which i eliminate using the group by in the SQL)
- smpa01
Community Champion
JacobCoops can you try this
Measure = CALCULATE(DISTINCTCOUNT('Table'[storeNumber]),ALLEXCEPT('Table','Table'[orderNumber]))
- CNENFRNL
Community Champion
SELECT [OrderNr] , [LagerNr] , COUNT([LagerNr]) OVER (PARTITION BY [LagerNr]) [#DIST] FROM XXX WHERE [OrderNr] = '12345' GROUP BY [OrderNr] , [LagerNr]