Forum Discussion

AnGorSan's avatar
AnGorSan
Frequent Visitor
8 years ago

Issue with visual level filters and ParallelPeriod Function (MDX)

Hello Comunity,   We use live connection to MOLAP Cube, and have the strange issue with visual level filters and previous periods based data. In the cube the measure is defined as CREATE MEMBER C...
Need Help
v-huizhn-msft's avatar
v-huizhn-msft
Microsoft Employee
8 years ago

Hi AnGorSan,

When you use Direct Query, Power BI Desktop queries the underlying data source, which means you’re always viewing current data. You said "the Power BI visuals seem to filter the datasets underneath them", in your dataset, the data is not filtered?

 

Thanks,
Angelia

AnGorSan's avatar
AnGorSan
Frequent Visitor
8 years ago

Hi Angela,

 

I do not understand why you are using the term "Direct Query".

"Direct Query" is the connection mode to the SQL databases.

 

We are using "LIVE CONNETCION" connection mode, to connect to the ANALYTICAL database, which is MOLAP cube.

I am saying, again, that our initial dataset is not filtered, and provides data for multiple years.

In my dataset there are measures that are calculated using MDX function PARALLELPERIOD.

So when I use a VISUAL filter, I expect it to be applicable to the data it should display, not to the query that is being sent to SSAS cube.

 

However, instead, the filter is applied to the DATASET, making the MDX function PARALLELPERIOD ineffective for the first displayed period, because it filters out the value required for the calculation!

 

  • v-huizhn-msft's avatar
    v-huizhn-msft
    Microsoft Employee
    8 years ago

    Hi AnGorSan,

    In your scenario, the filter is applied to the DATASET, so the date column used in PARALLELPERIOD are still continuous? The date should form a contiguous interval, please review to this thread.

     

    Best Regards,
    Angelia

    • AnGorSan's avatar
      AnGorSan
      Frequent Visitor
      8 years ago

      Hi!

       

      The problem we have in is exactly the same as mentioned at the end of the thread you proposed to look at.

      Adding extra month to dataset (as mentioned in a thread) is not a solution! its a workaround.

       

      Ok, I understand that the implementation of Power BI visual filters is to propagate the filters to dataset/query ALWAYS.

      This is misleading as the VISUAL filter implied the data is filtered on VISUAL level, and we, as developers, must come up with stupid workarounds, like get extra month data and hide empty values, instead of just applying visual filter.

       

      Quite far from what one would expect from a company like MS :(