Forum Discussion

kinkate18nic's avatar
kinkate18nic
Frequent Visitor
3 years ago
Solved

Count unique values present across months

Hello,

I have data set similar to below:

Event DateAlertID
1/20/202364hytv57684
1/20/202364hytv57684
2/20/2023hf6474jdd6
2/20/2023hyalam90yd
2/20/2023hyalam90yd
3/20/2023njs7109jdd6
4/20/2023njs7109jdd6
4/20/2023axm86djdkf
4/20/2023axm86djdkf

I want to count unique alertIDs month wise.

the measure : UniqueAlertIDCount = DISTINCTCOUNT('TableName'[AlertID])
counts alertID twice if its present in two months, in this case "njs7109jdd6" is present in both March and April, the out put i want is 

Month Unique AlertID Count

January1
February2
March1
April1

However alerting and modify measures, I still get: April as 2

"DAX"
"MEASURE"
  • tamerj1's avatar
    tamerj1
    3 years ago

    kinkate18nic 
    Yes, it will be counted as 1. But you can filter it out using either of the folowing depending on whether the blank is acually BLANK () or just empty string ""

    Unique AlertID Count =
VAR CurrentDate =
    MIN ( 'Table'[Event Date] )
VAR CurrentIDs =
    FILTER ( VALUES ( 'Table'[AlertID] ), 'Table'[AlertID] <> BLANK () )
VAR PreviousIDs =
    DISTINCT (
        SELECTCOLUMNS (
            FILTER ( ALLSELECTED ( 'Table' ), 'Table'[Event Date] < CurrentDate ),
            "AlertID", 'Table'[AlertID]
        )
    )
RETURN
    COUNTROWS ( EXCEPT ( CurrentIDs, PreviousIDs ) )


Unique AlertID Count =
VAR CurrentDate =
    MIN ( 'Table'[Event Date] )
VAR CurrentIDs =
    FILTER ( VALUES ( 'Table'[AlertID] ), 'Table'[AlertID] <> "" )
VAR PreviousIDs =
    DISTINCT (
        SELECTCOLUMNS (
            FILTER ( ALLSELECTED ( 'Table' ), 'Table'[Event Date] < CurrentDate ),
            "AlertID", 'Table'[AlertID]
        )
    )
RETURN
    COUNTROWS ( EXCEPT ( CurrentIDs, PreviousIDs ) )

     

     

4 Replies

  • tamerj1's avatar
    tamerj1
    Icon for Community Champion rankCommunity Champion
    3 years ago

    Hi kinkate18nic 

    Please try

    Unique AlertID Count =
    VAR CurrentDate =
    MIN ( 'Table'[Event Date] )
    VAR CurrentIDs =
    VALUES ( 'Table'[AlertID] )
    VAR PreviousIDs =
    DISTINCT (
    SELECTCOLUMNS (
    FILTER ( ALLSELECTED ( 'Table' ), 'Table'[Event Date] < CurrentDate ),
    "AlertID", 'Table'[AlertID]
    )
    )
    RETURN
    COUNTROWS ( EXCEPT ( CurrentIDs, PreviousIDs ) )

    • kinkate18nic's avatar
      kinkate18nic
      Frequent Visitor
      3 years ago

      Thank for the measure, this seems to be working fine. Just a hiccup, suppose if there is a blank alertID in a month, will that cause the count of the month to increase? I was testing this out in my dataset, for march i have actual 4 unique alertids and there is one entry with eventid but alert id blank. for march this measure is providing 5 as the count.

      • tamerj1's avatar
        tamerj1
        Icon for Community Champion rankCommunity Champion
        3 years ago

        kinkate18nic 
        Yes, it will be counted as 1. But you can filter it out using either of the folowing depending on whether the blank is acually BLANK () or just empty string ""

        Unique AlertID Count =
VAR CurrentDate =
    MIN ( 'Table'[Event Date] )
VAR CurrentIDs =
    FILTER ( VALUES ( 'Table'[AlertID] ), 'Table'[AlertID] <> BLANK () )
VAR PreviousIDs =
    DISTINCT (
        SELECTCOLUMNS (
            FILTER ( ALLSELECTED ( 'Table' ), 'Table'[Event Date] < CurrentDate ),
            "AlertID", 'Table'[AlertID]
        )
    )
RETURN
    COUNTROWS ( EXCEPT ( CurrentIDs, PreviousIDs ) )


Unique AlertID Count =
VAR CurrentDate =
    MIN ( 'Table'[Event Date] )
VAR CurrentIDs =
    FILTER ( VALUES ( 'Table'[AlertID] ), 'Table'[AlertID] <> "" )
VAR PreviousIDs =
    DISTINCT (
        SELECTCOLUMNS (
            FILTER ( ALLSELECTED ( 'Table' ), 'Table'[Event Date] < CurrentDate ),
            "AlertID", 'Table'[AlertID]
        )
    )
RETURN
    COUNTROWS ( EXCEPT ( CurrentIDs, PreviousIDs ) )