Forum Discussion

ChrisWilliams's avatar
ChrisWilliams
Advocate II
9 years ago
Solved

Rankx with filter

I'm working on a RANKX function that involves a filter.  I reviewed this answer, but it did not seem to help my scenario.

 

I have a table called "UserMessages", and a simplified version is shown below:

 

 

Basically, I want to rank users as "Top Senders" or "Top Receivers".  When they are a Top Sender, we are ranking based on filtering the Direction column to "Outbound".

 

I created a message called "Rank Top Senders" and it looks like this:

 

Rank Top Senders = RANKX(FILTER(ALL(UserMessages), UserMessages[Direction] = "Outbound"), SUM(UserMessages[Messages]),,DESC,Dense)

 

However, it doesn't appear to rank correctly, because each user is shown with a rank of 1

 

I've tried several variations, but if someone can point me in the right direction I'd appreciate it.

Need Help
  • Vvelarde's avatar
    Vvelarde
    9 years ago

    hi ChrisWilliams

     

    Rank Top Senders =
RANKX (
    FILTER ( ALL ( UserMessages ); UserMessages[Direction] = "Outbound" );
    CALCULATE (
        SUM ( UserMessages[Messages] );
        ALLEXCEPT ( UserMessages; UserMessages[User] )
    );
    ;
    DESC;
    DENSE
)

     

    replace ; with ,

7 Replies

  • Vvelarde's avatar
    Vvelarde
    Community Champion
    9 years ago

    hi ChrisWilliams

     

    Rank Top Senders =
RANKX (
    FILTER ( ALL ( UserMessages ); UserMessages[Direction] = "Outbound" );
    CALCULATE (
        SUM ( UserMessages[Messages] );
        ALLEXCEPT ( UserMessages; UserMessages[User] )
    );
    ;
    DESC;
    DENSE
)

     

    replace ; with ,

    • ChrisWilliams's avatar
      ChrisWilliams
      Advocate II
      9 years ago

       Thank you Vvelarde, that works perfectly.  I'll have to study up on AllExcept... I haven't seen that used in Ranking before but it works well.

    • Anonymous's avatar
      Anonymous
      Not applicable
      7 years ago

      Hi Chris,

       

      I have been searching for help on my challenge and run into this post. I tried to modify your solution to my situation but it is not working somehow. My situation is a little bit different.

       

      A simplified table is below. What I want is to rank the Customer Name by Sales BUT EXCLUDING all the Customers with Blank names (let's call them anyomymous customers). 

       

      The formula I came up after modelling yours is:

      Rank Customer By Sales= RANKX(
      filter(all(Table[Customer Name]),NOT(ISBLANK(Table[Customer Name]))),
      CALCULATE(sum(Sales), ALLEXCEPT(Customer Name])),,DESC,Dense)
       
      But it is still including the Anonymous customers. 

       

      Thanks for help. 

       

      PBISean

       

      Customer NameSalesSales Order Num
      A8001
      B7002
      C6003
      D5004
      E4005
       3006
      F2007
      G1008
      A809
      B7010
      C6011
      D5012
      E4013
       3014
      F2015
      G1016
      • Vvelarde's avatar
        Vvelarde
        Community Champion
        7 years ago

        Anonymous

         

        Hi, a simple way is using a measure to Rank (Basic) and a Visual Level Filter to exclude Blanks.

         

        If you have problems with the filter you can create a new measure

         

        MeasuretoexcludeAnonymus=Len(Table1[ClientName])

         

        and use it in the visual level filter to exclude the 0.

         

        Regards

         

        Victor