Forum Discussion

Necrosis1968's avatar
Necrosis1968
Regular Visitor
2 years ago
Solved

Filtering columns in a Matrix

Hi, I'm a Power BI Newbie so sorry for the simplicity of my question. I've looked on the forums but can't find anything that I can get to work.   I have a matrix based on a table. One column is cal...
Need Help
  • Sergii24's avatar
    Sergii24
    2 years ago

    Hi Necrosis1968, I assume you've built your matrix by dragging and dropping columns and then changing calcualtion to "Count (Distinct)". Is it correct?

    If so, you're using "implicit" measures as commented by Greg_Deckler. "Implicit" means that PowerBI makes a calculation for you and you don't need to write any formula.

    The reason you get 1 when filtering "Resolved" is because when apply it, your table is filtered only to 2 rows (Resolved = Yes) and each "Problem Type" has only 1 "Unique ID" at that moment. 

    What you might want to do is to "ignore" filter on "Resolved". To do so, you need to use "explicit" measures. The following measure will help you to achieve it:

     

    Count Unique ID = 
    CALCULATE(                                      //changes calculation context
        DISTINCTCOUNT( 'Table'[ UniqueID ] ),       //perform this operation    
        ALL( 'Table'[ Resolved] )                   //but only when this rule is applied (nevertheless you see it a row below it applies before  )
    )

     

    Now, the filter "Resolved" is ignored so you get a proper count. 


    You can use the similar logic to calcualte resolved, unresolved cases as well as find the resolution percentage:

     

    Count Resolved = 
    CALCULATE(
        DISTINCTCOUNT( 'Table'[ UniqueID ] ),
        'Table'[ Resolved] = "Yes"
    )

Count Not Resolved = 
    CALCULATE(
        DISTINCTCOUNT( 'Table'[ UniqueID ] ),
        'Table'[ Resolved] = "No"
    )

% Resolved = DIVIDE( [Count Resolved], [Count Unique ID] )

     

     
    Here is the final output:

    I'd suggest you to complete this course on YouTube before you going deeper in using PowerBI, it will help you a lot! Introducing DAX 101 videos (youtube.com)

    Good luck 🙂

Necrosis1968's avatar
Necrosis1968
Regular Visitor
2 years ago

I was using count, rather than count distinct on Resolved. Otherwise, when we use more rows in the sample data, we would end up with 1 as the count as Resolved can only be Yes or No. 

Thus a larger data set would look like:

Problem Type       UniqueID     Resolved

App                             A1               Yes

App                             A2               No

App                             A6               No

App                             A7               No

OS                               A3               No

OS                               A4               Yes

OS                               A8               Yes

 

With the desired output:

Problem Type          Count of Unique ID          Count of Resolved

App                                       4                                 1

OS                                         3                                 2

 

I tried the code you left, but it still doen't filter the Resolved column. I must be doing something fundamentally wrong.

 

Thanks for helping on the % which I'll attempt once the filter is working.

 

I'll start the YouTube link you suggested now! Thanks so much for the recommendation.

Necrosis1968's avatar
Necrosis1968
Regular Visitor
2 years ago

Also tried the % code but I get the error: 

The syntax for 'Count' is incorrect. (DAX( CALCULATE( DISTINCTCOUNT( 'Table'[UniqueID] ), 'Table'[Resolved] = "Yes" )Count Not Resolved = CALCULATE( DISTINCTCOUNT( 'Table'[UniqueID] ), 'Table'[Resolved] = "No" )% Resolved = DIVIDE( [Count Resolved], [Count Unique ID] )))

 

I have changed my table name to Table so it all matches yours.

 

Not that the filtering works though.

 

Silly really as in excel you use a pivot table and add Resolved as a filter. That's what I'm trying to replicate.

  • Sergii24's avatar
    Sergii24
    Super User
    2 years ago

    Hi Necrosis1968, the block of my previous reply contains 3 separate measures 😉

    Count Resolved = 
    CALCULATE(
        DISTINCTCOUNT( 'Table'[ UniqueID ] ),
        'Table'[ Resolved] = "Yes"
    )

     

    Count Not Resolved = 
    CALCULATE(
        DISTINCTCOUNT( 'Table'[ UniqueID ] ),
        'Table'[ Resolved] = "No"
    )

     

    % Resolved = DIVIDE( [Count Resolved], [Count Unique ID] )
    • Sergii24's avatar
      Sergii24
      Super User
      2 years ago

      P.S. some column names in my code have spaces in their names. Make sure to adjust them to your data model 😉

      • Necrosis1968's avatar
        Necrosis1968
        Regular Visitor
        2 years ago

        Thank you Sergii24 I now have it fully working. I have removed the columns and put the measures in and it works exactly as you have shown it!!!

    • Necrosis1968's avatar
      Necrosis1968
      Regular Visitor
      2 years ago

      Wow! Thank you that explains it!

  • Necrosis1968's avatar
    Necrosis1968
    Regular Visitor
    2 years ago

    Update! I have managed to get some success. I had to put the measure in the matrix values and remove the original unique ID I was using (hey I said I was a newb). The output now shows below. Not as neat as Sergii24 but getting there. Just can't get the syntax right of the % calculation in Sergii24 second piece of code. Thanks so much for your help so far both!