To calcuate the accumulated value

Question

Hello,
I would like to write a calc script for the following scenario: I have a series of data value, and would like to calcuate the accumulated value based on the sequence of from the smallest to the largest. The case is like this:
The input is as following
            Case 1
Acct1: 10
Acct2:    5
Acct3:    8
Acct4: 15
I would like to calcuate the accumlated value based on the data value, the caulcated result is at column Case 2:
           Case 1  Case 2  Rank
Acct1: 10          23           3
Acct2:    5           5           1
Acct3:    8          13          2
Acct4: 15              38       4
I can easily use the rank function to rank the data value of "Case1", but after that how can I use this rank to cauculate the accumluated values? The data value may change, (from time to time, for example).
Any good suggestions? Thanks in advance

Answer 1

Hi There
Try this script below. I have done this in the SAMPLE app. script may look crazy, but I have tested this and it works. Also what makes it a little more complicated is the fact that ranking can give you duplicate rank values, so I have designed this to get around that problem.
Thanks
Anthony
SET UPDATECALC OFF;
VAR AccountCount=0;
VAR CurrRank=0
/*Create blocks in case dimension as mine is sparse*/
FIX(Local,FY13,BegBalance,Forecast,Working,"E01_101_1110",@RELATIVE("TotaltestAccount",0),"HSP_InputValue")
DATACOPY "Case1" TO "Rank";
DATACOPY "Case1" TO "Case2";
DATACOPY "Case1" TO "Sequence";
DATACOPY "Case1" TO "Case1Seq";
ENDFIX
FIX(Local,FY13,BegBalance,Forecast,Working,"E01_101_1110",@RELATIVE("TotaltestAccount",0),"HSP_InputValue")
"Case2" = #missing;
"Sequence"= #missing;
/*Create a sequence for the list of accounts so that if you for any chance have a value that is the same the rank will give a unique rank value - because this becomes a problem later*/
"Sequence"= (@SHIFT("Sequence",-1,"Acc1":"Acc5") + 0.00000000001);
/*Rank case1 and add the sequence to give you the unique ranking and then * -1 to reverse the rank*/
"Case1Seq" = "Sequence" + @RANK(SKIPNONE,"Case1",@RANGE("Case1",@RELATIVE("TotaltestAccount",0))) * -1;
ENDFIX
FIX(Local,FY13,Forecast,Working,"E01_101_1110",@RELATIVE("TotaltestAccount",0),"HSP_InputValue")
/*Next rank the previous rank from above to now populate the unique ranking of case 1 in dec order*/
"Rank"(
      IF(@ISMBR("BegBalance"))
"Rank" = @RANK(SKIPNONE,"Case1Seq",@RANGE("Case1seq",@RELATIVE("TotaltestAccount",0)));
ENDIF
/*Populate the max ranking value to start the variable counter*/
AccountCount = @MAXSRANGE(SKIPMISSING,"Rank"->"BegBalance",@RELATIVE("TotalTestAccount",0));
CurrRank = AccountCount;
ENDFIX
FIX(Local,FY13,"Case2",Forecast,Working,"E01_101_1110",@RELATIVE("TotaltestAccount",0),"HSP_InputValue")
/*Calculate the lowest case 1 value first or highest rank value, then make curr rank equal to highest rank -1 */
"BegBalance"(
           IF(AccountCount == "Rank")
           "Begbalance" = "BegBalance"->"Case1" ;
               CurrRank = CurrRank - 1;
           ENDIF
/*Then loop through the process reducing the current rank variable by 1 each time*/
LOOP(100)
"BegBalance"(
         IF(CurrRank == "Rank")/* i.e rank ==4 */
      "Begbalance" = @MAXSRANGE(SKIPMISSING,"Case2"->"BegBalance",@RELATIVE("TotalTestAccount",0)) +"BegBalance"->"Case1";
          CurrRank = CurrRank - 1;/*then make rank == 3*/
           ENDIF)
ENDLOOP;
ENDFIX