Forum Discussion

JustinDoh1's avatar
JustinDoh1
Post Prodigy
5 years ago
Solved

Exclude distinct count with OR

I am trying to write DAX formula with following logic:

 

Count distinct "ClientID" if one has "Refused" in Consent column,

BUT, if one has "Consented" in Consent column, do not count (exclude) distinct "ClientID". 

 

So, client 69 and 12468 should be excluded from counting.

 

This is my original DAX:

 

Refused = CALCULATE (

                                          DISTINCTCOUNT( Table[ClientID] ),
                                          Table[Consent] = "Refused"
                                       )
 ---------------------------------------------------------------------------
Do  I just subtract like this?
 

Refused = CALCULATE (

                                           ( DISTINCTCOUNT(Table[ClientID] ),
                                              (
                                                   ( Table[Consent] = "Refused" )
                                                     -  
                                                    (Table[Consent] = "Refused" && Table[Consent] = "Consented" )

                                              )

                                        )

I don't think it works.

I need to find the way to calculate the count ClientID level, not table level. 

 

 

Help with DAX
  • Jos_Woolley's avatar
    Jos_Woolley
    5 years ago

    Ok, thanks for clarifying.

    Refused Not Consented =
VAR Refused =
    DISTINCT (
        SUMMARIZE (
            FILTER ( 'Table', 'Table'[Consent] = "Refused" ),
            'Table'[ClientID]
        )
    )
VAR Consented =
    DISTINCT (
        SUMMARIZE (
            FILTER ( 'Table', 'Table'[Consent] = "Consented" ),
            'Table'[ClientID]
        )
    )
RETURN
    COUNTROWS ( EXCEPT ( Refused, Consented ) )

    Regards

7 Replies

  • Jos_Woolley's avatar
    Jos_Woolley
    Solution Sage
    5 years ago

    Ok, thanks for clarifying.

    Refused Not Consented =
VAR Refused =
    DISTINCT (
        SUMMARIZE (
            FILTER ( 'Table', 'Table'[Consent] = "Refused" ),
            'Table'[ClientID]
        )
    )
VAR Consented =
    DISTINCT (
        SUMMARIZE (
            FILTER ( 'Table', 'Table'[Consent] = "Consented" ),
            'Table'[ClientID]
        )
    )
RETURN
    COUNTROWS ( EXCEPT ( Refused, Consented ) )

    Regards

    • JustinDoh1's avatar
      JustinDoh1
      Post Prodigy
      5 years ago

      Jos_Woolley  Thank you for your help. I am trying to understand the logic here. First, we are  counting all rows that have a word "Refused". Then, we are also counting all rows that have a word "Consented". Is it right? Then, what does the rest of statement mean? I think I understand what 'Except' means, but from what dataset (from all criteria - including others ('Historial', 'Not Eligible', etc.)? Or am I totally off the track? Thank you.

      • Jos_Woolley's avatar
        Jos_Woolley
        Solution Sage
        5 years ago

        The first part:

        VAR Refused =
DISTINCT (
    SUMMARIZE (
        FILTER ( 'Table', 'Table'[Consent] = "Refused" ),
        'Table'[ClientID]
    )
)

        defines the variable 'Refused' as the single-column table comprising the distinct values from the ClientID column for which the Consent column entry is "Refused".

        The next variable is similarly defined, though for Consent column entries of "Consented".

        The EXCEPT clause then returns a single-column table comprising all Client ID entries from the 'Refused' table which do not appear in the 'Consent' table.

        Finally, the number of rows in this last table are counted.

        Regards

         

  • Jos_Woolley's avatar
    Jos_Woolley
    Solution Sage
    5 years ago

    Hi,

    In the example you give, the only two possible entries in the 'Consent' column are 'Consented' and 'Refused'. As such, you can use:

    Refused =
DISTINCTCOUNT ( 'Table'[ClientID] )
    - CALCULATE (
        DISTINCTCOUNT ( 'Table'[ClientID] ),
        'Table'[Consent] = "Consented"
    )

    If other entries are in fact possible within the 'Consent' column then please update your post with a more realistic dataset.

    Regards

    • JustinDoh1's avatar
      JustinDoh1
      Post Prodigy
      5 years ago

      Jos_Woolley Sorry. actually, there are different options for the value in the  "Consent" column (like 'Historial', 'Not Eligible').