Comp variables have prominent importance in all mainframe technologies COBOL/IMSDB/IMSDC/CICS.
The advantage is, at word boundaries it is more efficient than comp-3, means takes less space.
The popular range:
PIC 9(1) — PIC 9(4) 2 bytes
PIC 9(5) — PIC 9(9) 4 bytes
PIC 9(10) — PIC 9(18) 8 bytes
As we aware the above values only for “COMP” data type. Let us take on “COMP-1” , so the data will store internally as 1 word. It takes 4 bytes of space.
Then, for Comp-2, the data will store internally as 2 words. So it takes 8 bytes. This is the reason Comp-2 is more precession than Comp-1.
The other great point is, we should not specify PIC for Comp-1 and Comp-2, the reason is, the data type(numeric) and sizes (4 or 8 bytes) are pre-determined. So PIC is not required.
PIC clause determined relation between data name and data type.
Bottom Line: Comp-1 take 4 bytes(1 word) is single precession & Comp-2 takes 8-bytes(2 words) is double precession.
05 DATA-ONE USAGE IS COMP-1
05 DATA-TWO USAGE IS COMP-2