Issue with Reading data with FOR EACH LOOP.

Question

Hi Guys,
Scenario
We Plan Sales of an Article on a half years(fiscal variant MQ with 2 periods of half year each). We also plan the
number of Weeks in Sales as a Key figure.  The cummulative Sales for two half years in then distributed to number of weeks for which the Article is sold (0CALWEEK)
In another Cube we plan a distribution profile. This cube contains just three elements.
Characteristic = ZPROFILE
TIme Char = 0CALWEEK
Keyfigure = ZDIST
(Note : You will notice The distribution profile is not stored per Article here)
The logic of distribution is :
1.  Sales of Article is cummulated for 1 years ( eg. for periods 001.2010 & 002.2010 )
2.  Find the number of weeks in Sales for this Article ( eg. 14 weeks. ) [Technically we store this value on the first fiscal
period. In our example it is stored in 001.2010
3.  User choose a distribution profile in a variable ( e.g ZH55 )
4   ZDIST is summed for number of Sale Weeks for this Article. in our case it is14 weeks.
5. Distribution = Sales of Article (WEEK N) = [Sales Cummulated for 1 year] * [ZDIST(Week N) / ZDIST(sum of 24 weeks)]
This loop is run 14 times. (N from 201001 to 201014.)
TECHICAL DETAILS
Cube 1 (0CM_SKU = Material, 0CPBAINSABU = SALES key figure, WEEKINSALE = Number of weeks for which Material will be
sold). The Sales Planned data is stored as # for 0CALWEEK.
0CM_SKU/0CALWEEK/0FISCPER/0FISCVARNT/0CPBAINSABU--/WEEKINSALE
SHIRT/000000/2010001/ZP/100EUR--/24
SHIRT/000000/2010002/ZP/200EUR--/00
PANTS/000000/2010001/ZP/300EUR--/35
PANTS/000000/2010002/ZP/200EUR--/00
CUBE 2 ( the profile data is stored here )
ZPROFILE/0CALWEEK/ZDIST
ZH55/201001/10
ZH55/201002/12
ZH55/201003/8
ZH55/201004/11
ZH55/201005/10
..so on.
9:11 PM
Edited by: Dheeraj Gupta on Apr 5, 2010 9:20 PM

Answer 1

My Sample CODE IS BELOW.
DATA CW1 TYPE 0CALWEEK.
DATA CW2 TYPE 0CALWEEK.
DATA PER1 TYPE 0FISCPER.
DATA PER2 TYPE 0FISCPER.
DATA FVAR TYPE 0FISCVARNT.
DATA PROV1 TYPE 0INFOPROV.
DATA PROV2 TYPE 0INFOPROV.
DATA SKU TYPE 0CM_SKU.
DATA DPROF TYPE ZPROFILE.
DATA CUMSALES TYPE F.
DATA TEMP TYPE F.
DATA PRO TYPE F.
DATA COUNT TYPE I.
*I WILL READ THE PROFILE VALUE FROM VARIABLE IN ACTUAL PROGRAM.
HARDCODED FOR TEST PURPOSES. 
DPROF = 'ZH55'.
*FISCAL YR VARIANT
FVAR = 'MQ'.
*HARDCODED FOR TEST PURPOSES. 
PER1 = 2010001.
PER2 = TMVL(PER1, 1, FVAR).
*CUBE FOR SALES
PROV1 = 'ZCLARK'.
*CUBE FOR PROFILE
PROV2 = 'C_ZCLARK'.
FOREACH SKU.
*SUM UP SALES FOR ONE YEAR
*Operand:{Key figure name, 0CALWEEK, 0CM_SKU, 0FISCPER, 0FISCVARNT, 0INFOPROV, ZPROFILE}
CUMSALES = CUMSALES + {0CPBAINSABU,#,SKU,PER1, FVAR,PROV1,#} + {0CPBAINSABU,#,SKU,PER2, FVAR,PROV1,#}.
*READ NUMBER OF WEEKS OF SALES
COUNT = {ZNWEEK,#,SKU,PER1, FVAR,PROV1,#}.
*HARDCODED FOR TEST PURPOSES. 
Starting week of distrbution is first week of the year
CW2 = 201001.
CW1 = 201001.
*HELP NEEDED HERE ***********
*NOT ABLE TO READ THIS PROFILE DATA
DO COUNT TIMES.
   PRO = PRO + {ZPROF,CW2,#,#,#,PROV2,ZPROFILE}.
   CW2 = TMVL(CW2, 1).
ENDDO.
DO COUNT TIMES.
{0CPBAINSABU,CW1,SKU,#, #,PROV1,#} = CUMSALES * {ZPROF,CW1,#,#, #,PROV2,ZPROFILE} / PRO.
CW1 = TMVL(CW1, 1).
ENDDO.
*RESET
CUMSALES = 0.
ENDFOR.