Forum Discussion

cursty's avatar
cursty
Frequent Visitor
1 year ago
Solved

Calculations with virtual table

Hello everyone,   I am working with expenses dataset. This dataset contains 4 columns: the unique id, item id (which can appear multiple times in case few payments have been made for the same item)...
Need Help
  • Anonymous's avatar
    Anonymous
    1 year ago

    Thanks for the reply from lbendlin and Greg_Deckler, please allow me to provide another insight.
    Hi cursty ,

    Please try the following dax.

    Perc_Rank = 
VAR _tb = CALCULATETABLE(SUMMARIZECOLUMNS('Expenses'[item_id],"Expense",[Expenses]),ALL())
VAR _expenses = VALUE([Expenses])
VAR _lower = IF(COUNTROWS(FILTER(_tb, [Expense]<_expenses))=BLANK(),0,COUNTROWS(FILTER(_tb, [Expense]<_expenses)))
VAR _higher = IF(COUNTROWS(FILTER(_tb, [Expense]>_expenses))=BLANK(),0,COUNTROWS(FILTER(_tb, [Expense]>_expenses)))
RETURN
DIVIDE(_lower,_higher+_lower,0)


    The result is as follows, hopefully it will meet your needs.


    Please see the attached pbix for reference.

    Best Regards,
    Dengliang Li

    If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

lbendlin's avatar
lbendlin
Super User
1 year ago

What is your expected result and why are you not using the RANK function?

cursty's avatar
cursty
Frequent Visitor
1 year ago

I overcomplicated this formula. I was trying use RANKX, but I've been getting 1 for every item on the list. I am actually looking for the equivalent of the Excel PERCENTRANK.INC formula. My expected output is shown below:

 

Item IDAmountPerc
item-1130050%
item-22000%
item-34500100%
  • lbendlin's avatar
    lbendlin
    Super User
    1 year ago

    Greg_Deckler Did you cover that in your mapping table?

      • cursty's avatar
        cursty
        Frequent Visitor
        1 year ago

        Thank you for suggestion Greg_Deckler lbendlin 

         

        I used your measure in my dashboard, but unfortunately it still doesn't resolve my problem as I'm receiving value 0 for the item with highest expenses and blanks for the rest of them.

         

        Below, I highlighted what I've changed in your original code

        PERCENTILERANK.INC =
        VAR maxdt = MAX('Calendar'[Date])
        VAR mindt = EDATE(maxdt, -36) + 1
        VAR itm = ADDCOLUMNS(VALUES(Expenses[item_id]), "tv", CALCULATE(SUM(Expenses[paid_amount]), Expenses[payment_date] IN CALENDAR(mindt, maxdt)))
        VAR __Value =  MAXX(SUMMARIZE(Expenses, Expenses[item_id], "Amount", SUM(Expenses[paid_amount])), [Amount])


        RETURN
            IF(
                __Value IN SELECTCOLUMNS(itm,"Values",[tv]),
                        VAR __NumLower = COUNTROWS(FILTER(itm,[tv] < __Value))
                        VAR __NumHigher = COUNTROWS(FILTER(itm, [tv] > __Value))
                        VAR __Rank = __NumLower / (__NumLower + __NumHigher)
                    RETURN IF(ISBLANK(__Rank),0,__Rank),
                        VAR __Lower = MAXX(FILTER(itm,[tv] < __Value),[tv])
                        VAR __Higher = MINX(FILTER(itm,[tv] > __Value),[tv])
                        VAR __LowerNumLower = COUNTROWS(FILTER(itm,[tv] < __Lower))
                        VAR __LowerNumHigher = COUNTROWS(FILTER(itm,[tv]> __Lower))
                        VAR __LowerRank = __LowerNumLower / (__LowerNumLower + __LowerNumHigher)
                        VAR __HigherNumLower = COUNTROWS(FILTER(itm,[tv]< __Higher))
                        VAR __HigherNumHigher = COUNTROWS(FILTER(itm,[tv] > __Higher))
                        VAR __HigherRank = __HigherNumLower / (__HigherNumLower + __HigherNumHigher)
                    RETURN
                        __LowerRank + ( __Value - __Lower ) / (__Higher - __Lower ) * ( __HigherRank - __LowerRank )
            )
    • cursty's avatar
      cursty
      Frequent Visitor
      1 year ago

      I am not sure if I understood your question, but for this measure I didn't use any mapping besides Calendar Table, which is needed to define the dates range.

       

      Below is more simple formula with RANKX function (user for var 'ranking' which is returning me 1 for every item on the list)

      Perc_Rank2 =
      var maxdt = max('Calendar'[Date])
      var mindt = edate(maxdt,-36)+1
      var items =  SUMMARIZE(FILTER(Expenses, Expenses[payment_date] IN CALENDAR(mindt,maxdt)), Expenses[item_id], "Value", SUM(Expenses[paid_amount]))
      var currentvalue =  CALCULATE(SUM(Expenses[paid_amount]), Expenses[payment_date] IN CALENDAR(mindt, maxdt))
      var total = COUNTROWS(DISTINCT(SELECTCOLUMNS(FILTER(ALLSELECTED(Expenses), Expenses[payment_date] IN CALENDAR(mindt, maxdt)), "UniqueID", Expenses[item_id])))
      var ranking = RANKX(items, currentvalue,,ASC)
      RETURN DIVIDE(ranking, total)*100