Quetion on "UNIT_CONVERSION_WITH_FACTOR"

Question

Hi All
The function module "UNIT_CONVERSION_WITH_FACTOR" has been used in one of the standard cubes 0SD_C03.
This function is being called from a routine defined against the key figure 0QUANT_B.
Now my doubt is in the function module "UNIT_CONVERSION_WITH_FACTOR", there are statements like
  DATA: t_output TYPE c,
        d_output TYPE i.
  DESCRIBE FIELD output TYPE     t_output
                        DECIMALS d_output.
  IF t_output = 'P'.
    CALL FUNCTION 'ROUND'
      EXPORTING
        decimals      = d_output
        input         = result
        sign          = 'X'
      IMPORTING
        output        = output
      EXCEPTIONS
        input_invalid = 1
        overflow      = 2
        type_invalid  = 3
        OTHERS        = 4.
In the above piece of code "t_output" has not been assigned to any field in the COMM_STR.
Then how is it validating against some value in the statement "IF t_output = 'P'" ?
Can anybody tell me if "t_output" is predefined or pre-populated?
These things are from the Business Content point of view. No customizations done.
Hope I am clear and the question makes sense.
Thanks In Advance

Answer 1

sorry, I missed your question,
the d_output will hold the number of decimals.
So if your taget KF as data type packet (17, 3), meaning length 17 and 3 decimals
t_output will be P
d_output will be 3.
The subsequent rounding will be made with 3 decimals.
This is because the calculation is posted to a number of type FLOAT:
DATA result TYPE f.
  result = ( numerator * input ) / denominator + add_const. "B20K079016
This number has to be rounded with the same decimals than your target KF.
hope it is clearer now,
Olivier.