11 Replies Latest reply on Jun 15, 2016 11:33 AM by savas.ziplies

    Building a unique Count Distinct over all values down a dimension (unique running count)




      I face the following scenario:

      I have multiple data sources, everything is filtered, I result in a set of leading values (keys). Each key has an amount of data in another data source I would like to count distinct. That works, as far as I can tell.


      What I would like to have is a unique count, meaning the data in the other data source should only be counted once in the whole calculation.


      So, for example I have a filtered key list of

      * Key A

      * Key B

      * Key C


      Each key has some part numbers in the other data source, and some of these might be overlapping. In total, I only want to have them counted once. So, if the same part number is on Key A and Key B, the total count should be 1.


      How could this be achieved?



      Thanks a lot in advance