Problem:

Is this working correctly

WS-UNSIGNED    PIC 9(15)    COMP-5 VALUE ZERO.

WS-SIGNED        PIC S9(15)  COMP-5 VALUE ZERO.

WS-BYTES         PIC X(20)    VALUE '-1380985813'

MOVE FUNCTION NUMVAL(WS-BYTES) TO WS-UNSIGNED

MOVE FUNCTION NUMVAL(WS-BYTES) TO WS-SIGNED

After the above operation was performed:

WS-UNSIGNED = 2057592656942123

WS-SIGNED = -0000001380985813

Can this be explained?

Resolution:

The explanation has to do with the definition of the compiler directive comp-5.

COMP-5 compiler directive  by default is set to COMP-5"2"

COMP-5(2) says that negative numbers are stored in an unsigned COMP-5 item in 2's complement form. This is exactly what is happening.

If you want the sign preserved then you need to store the value in a signed data item.

If you want the absolute value then use ABS intrinsic functions in conjunction with NUMVAL

MOVE FUNCTION ABS(FUNCTION NUMVAL(WS-BYTES)) TO WS-UNSIGNED MOVE

If you change the COMP-5 directive to COMP-5"1" then the results will show the correct decimal value without the leading minus sign.

from the on-line docs on COMP-5 compiler directive:

COMP-5

Specifies whether the sign is to be dropped when a value is stored in an unsigned COMP-5 data item.

Syntax:

>>-.---.--COMP-5--"integer"----------------><

    -/-

Parameters:

integer Must be one of:

1 Behavior as in earlier versions of this Compiler. The sign is dropped.

2 The sign is not dropped. Negative numbers are stored in two's complement form, so that, except for their byte-order being machine dependent, unsigned COMP-5 items behave like COMP-X. This results in highly efficient arithmetic on unsigned COMP-5 items.