Forum Discussion

Nevos's avatar
Nevos
Frequent Visitor
6 years ago

Inconsistant filter results

I'm getting some really odd behaviour when filtering tables using a measure.  I have two tables (one a copy of the other) and when i remove two collumns from the second table my results change. backg...
General Comment
v-alq-msft's avatar
v-alq-msft
Community Support
6 years ago

Hi, Nevos 

 

There are two ways. One way is to make the fields in the visual 'Don't summarize'. Another way is to create a new measure as follows.

 

TotalInbound = CALCULATE(SUM(RealTimeData[InboundCalls]),FILTER('RealTimeData','RealTimeData'[LastUpdated] = [LatestDate]))

 

 

Result:

 

Best Regards

Allan

 

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

 

Nevos's avatar
Nevos
Frequent Visitor
6 years ago

Hi Allan and thanks for the response.  Yes removing the field to 'Don't summerise' gives the desired result but it doesn't explain why?  Both tables have the field set to SUM and are referencing the same data, measures and fields.  So why two different results?

 

I've tried your suggestion of the new measure but it doesn't give any data at all 😞 

  • tex628's avatar
    tex628
    Community Champion
    6 years ago

    Is the 3rd field, "Internal calls" also set to sum? 

    • Nevos's avatar
      Nevos
      Frequent Visitor
      6 years ago

      No, if i change that to SUM that changes the values in ALL other columns to give the SUM of all data and not just the most recent row.

       

      i don't understand that either 

      • tex628's avatar
        tex628
        Community Champion
        6 years ago

        Alright. The reason that your two tables are behabving differently is because "InternalCalls" is not set to be summarized.

        I will try to explain why. 

        When all of the columns are summarized there are no specific dates remaining to be used in the comparison of isLatest. You're expecting it to get the individual dates of the rows but since everything is summarized there are no rows to get the date from.


        Essentailly what this means is that your two measures ends up doing the same thing. In your first measure you use the ALLEXCEPT to ignore the individual rows and find the max date related to the specific agent. 

        In the second measure you end up doing the exact same thing with the max statement since it's evaluated over all of all the rows for each agent. 

        My advice would be to use a calulated column in the dataset to determine the most recent rows beforehand and then use that as the filter instead. 

        I hope this was some help! 

        Br,
        J