Forum Discussion

mscopuser's avatar
mscopuser
Frequent Visitor
2 years ago
Solved

Current Streak for Subject

Hi,   I have this table (see below). From that table I want to have calculation of current streak for doing a sub project. I have tried chat gpt but no luck.   Team member Client Projects...
Need Help
  • Greg_Deckler's avatar
    Greg_Deckler
    2 years ago

    mscopuser Sure, it's almost identical. Attached PBIX file below sig has the Current Streak measure as well as an example of Longest Streak for Project instead of sub-project.

    Current Streak Sub-Project = 
    VAR __SubProject = MAX( 'Table'[Sub-projects] )
    VAR __MaxDate = MAXX( ALL( 'Table' ), [Start date] )
    VAR __MinDate = 
        MINX( 
            FILTER( 
                ALL( 'Table' ), 
                [Sub-projects] = __SubProject 
            ), 
            [Start date] 
        )
    VAR __Table = 
        ADDCOLUMNS(
            CALENDAR( __MinDate, __MaxDate ),
            "__Count", 
                VAR __Date = [Date]
                VAR __Result = 
                    IF( 
                        COUNTROWS( 
                            FILTER( 
                                ALL( 'Table' ), 
                                [Start date] = __Date && [Sub-projects] = __SubProject 
                            ) 
                        ), 1, 0 
                    )
            RETURN
                __Result
        )
    VAR __Table1 = 
        ADDCOLUMNS(
            __Table,
            "Cthulhu", 
                VAR __Index = [Date]
                VAR __Group = [__Count]
                VAR __TmpTable1 = FILTER( __Table, [__Count] = __Group && [Date] < __Index)
                VAR __TmpTable2 = 
                    ADDCOLUMNS( 
                        __TmpTable1, 
                        "__diff", 
                        ( 
                            [Date] - 
                            MAXX( 
                                FILTER( 
                                    __Table, 
                                    [Date] < EARLIER( [Date] ) && [__Count] = EARLIER( [__Count] ) 
                                ), 
                                [Date] 
                            ) 
                        ) * 1. 
                    )
                VAR __Max = MAXX( __TmpTable2, [Date])
                VAR __MaxStart = MAXX( FILTER(__TmpTable2, [__diff] > 1), [Date])
                VAR __TmpTable3 = FILTER( __TmpTable2, [Date] >= __MaxStart)
                VAR __Result = 
                    SWITCH( TRUE(),
                        ISBLANK( __Max ), 1, 
                        __Max = __index - 1, COUNTROWS( __TmpTable3 ) + 1, 
                        1
                    )
            RETURN
                __Result
        )
    VAR __CurrentCount = MAXX( FILTER( __Table1, [Date] = __MaxDate ), [__Count] )
    VAR __Result = IF( __CurrentCount = 0, 0, MAXX( FILTER( __Table1, [Date] = __MaxDate ), [Cthulhu] ) )
RETURN
    __Result
Greg_Deckler's avatar
Greg_Deckler
Icon for Community Champion rankCommunity Champion
2 years ago

mscopuser Found this from the email you sent. I believe I solved this, see the attached PBIX below signature. 

Longest Streak = 
    VAR __SubProject = MAX( 'Table'[Sub-projects] )
    VAR __MinDate = MINX( FILTER( ALL( 'Table' ), [Sub-projects] = __SubProject ), [Start date] )
    VAR __MaxDate = MAXX( FILTER( ALL( 'Table' ), [Sub-projects] = __SubProject ), [Start date] )
    VAR __Table = 
        ADDCOLUMNS(
            CALENDAR( __MinDate, __MaxDate ),
            "__Count", 
                VAR __Date = [Date]
                VAR __Result = IF( COUNTROWS( FILTER( ALL( 'Table' ), [Start date] = __Date && [Sub-projects] = __SubProject ) ), 1, 0 )
            RETURN
                __Result
        )
    VAR __Table1 = 
        ADDCOLUMNS(
            __Table,
            "Cthulhu", 
                VAR __index = [Date]
                VAR __group = [__Count]
                VAR __tmpTable1 = FILTER( __Table, [__Count] = __group && [Date] < __index)
                VAR __tmpTable2 = ADDCOLUMNS( __tmpTable1, "__diff", ( [Date] - MAXX( FILTER( __Table, [Date] < EARLIER( [Date] ) && [__Count] = EARLIER( [__Count] ) ), [Date] ) ) * 1. )
                VAR __max = MAXX( __tmpTable2, [Date])
                VAR __maxStart = MAXX( FILTER(__tmpTable2, [__diff] > 1), [Date])
                VAR __tmpTable3 = FILTER( __tmpTable2, [Date] >= __maxStart)
                VAR __Result = 
                    SWITCH( TRUE(),
                        ISBLANK( __max ), 1, 
                        __max = __index - 1, COUNTROWS( __tmpTable3 ) + 1, 
                        1
                    )
            RETURN
                __Result
        )
    VAR __Result = MAXX( FILTER( __Table1, [__Count] <> 0 ), [Cthulhu] )
RETURN
    __Result
  • mscopuser's avatar
    mscopuser
    Frequent Visitor
    2 years ago

    You are life saver, I am so happy now. Chatgpt could not help me after spending 2 hours today.

     

    It's almost perfect, is see you have calculated the longest streak wich is data point I totally can use for me power bi dashboard I am trying to build.

     

    Could also you also provide one for the current streak? 

     

    I would be agian thank full 

     

    EDIT: I mean the current streak of each subproject. So if I would start to day it would be 1. If started a week ago it would be 7.

    • Greg_Deckler's avatar
      Greg_Deckler
      Icon for Community Champion rankCommunity Champion
      2 years ago

      mscopuser Sure, it's almost identical. Attached PBIX file below sig has the Current Streak measure as well as an example of Longest Streak for Project instead of sub-project.

      Current Streak Sub-Project = 
    VAR __SubProject = MAX( 'Table'[Sub-projects] )
    VAR __MaxDate = MAXX( ALL( 'Table' ), [Start date] )
    VAR __MinDate = 
        MINX( 
            FILTER( 
                ALL( 'Table' ), 
                [Sub-projects] = __SubProject 
            ), 
            [Start date] 
        )
    VAR __Table = 
        ADDCOLUMNS(
            CALENDAR( __MinDate, __MaxDate ),
            "__Count", 
                VAR __Date = [Date]
                VAR __Result = 
                    IF( 
                        COUNTROWS( 
                            FILTER( 
                                ALL( 'Table' ), 
                                [Start date] = __Date && [Sub-projects] = __SubProject 
                            ) 
                        ), 1, 0 
                    )
            RETURN
                __Result
        )
    VAR __Table1 = 
        ADDCOLUMNS(
            __Table,
            "Cthulhu", 
                VAR __Index = [Date]
                VAR __Group = [__Count]
                VAR __TmpTable1 = FILTER( __Table, [__Count] = __Group && [Date] < __Index)
                VAR __TmpTable2 = 
                    ADDCOLUMNS( 
                        __TmpTable1, 
                        "__diff", 
                        ( 
                            [Date] - 
                            MAXX( 
                                FILTER( 
                                    __Table, 
                                    [Date] < EARLIER( [Date] ) && [__Count] = EARLIER( [__Count] ) 
                                ), 
                                [Date] 
                            ) 
                        ) * 1. 
                    )
                VAR __Max = MAXX( __TmpTable2, [Date])
                VAR __MaxStart = MAXX( FILTER(__TmpTable2, [__diff] > 1), [Date])
                VAR __TmpTable3 = FILTER( __TmpTable2, [Date] >= __MaxStart)
                VAR __Result = 
                    SWITCH( TRUE(),
                        ISBLANK( __Max ), 1, 
                        __Max = __index - 1, COUNTROWS( __TmpTable3 ) + 1, 
                        1
                    )
            RETURN
                __Result
        )
    VAR __CurrentCount = MAXX( FILTER( __Table1, [Date] = __MaxDate ), [__Count] )
    VAR __Result = IF( __CurrentCount = 0, 0, MAXX( FILTER( __Table1, [Date] = __MaxDate ), [Cthulhu] ) )
RETURN
    __Result
      • mscopuser's avatar
        mscopuser
        Frequent Visitor
        2 years ago

        Thank you so much and for providing the Projects example. Al ready seeing great insights. This makes me want to learn power bi. Henche I will buy your book. 

         

        I am so happy.