COMP to Char

koulmahesh · New User Joined: 19 Jul 2010 Posts: 2 Location: Bangalore

Explain the output.

WORKING-STORAGE SECTION.

01 HEX-91-VALUE PIC 9(4) COMP VALUE 145.
01 HEX-91X REDEFINES HEX-91-VALUE.
15 FILLER PIC X(1).
15 HEX-91 PIC X(1).

PROCEDURE DIVISION.

DISPLAY "HEX-91-VALUE " HEX-91-VALUE
DISPLAY "HEX-91 " HEX-91

Output we got:

HEX-91-VALUE 0145
HEX-91 j
[/Search]

Binop B · Posted: Mon Jul 26, 2010 12:20 pm

Hi Mahesh,

Why dont you start with "What u understood so far ?"

Robert Sample · Posted: Mon Jul 26, 2010 4:35 pm

There is a link to manuals at the top of the page. Click on it, find the COBOL Programming Guide manual, and locate section 1.3.4.7 of this manual. You will find out how COMP variables are stored internally. Then go the COBOL Language Reference manual and read Appendix 3.1 on the EBCDIC collating sequence. This should allow you to explain the output yourself.

CICS Guy · Posted: Mon Jul 26, 2010 5:29 pm

dick scherrer · Posted: Mon Jul 26, 2010 8:06 pm

Hello and welcome to the forum,

koulmahesh · New User Joined: 19 Jul 2010 Posts: 2 Location: Bangalore

Hey .. cool ..

Thanks friends..

saurabh39 · Posted: Thu Jul 29, 2010 11:52 pm

One question though.......how will display different value when we are referring to same memory location?

Robert Sample · Posted: Fri Jul 30, 2010 12:09 am

Tushar, the different values displayed depend upon how you look at the data. For example, if you have

saurabh39 · Posted: Fri Jul 30, 2010 12:58 am

@Robert - Thanks for Quick Clarification.

But One more Clarification required- I understand the hex value for p & r is 97,99 and for left parenthesis 4D, so the value will be 97994D, since D is negative it will be -97994. Till this point I am clear, But you said it will be displayed as 9799M, but M is D4, but we are storing as 4D which is (, so the value as per my understanding should be 9997(.

AM I Nissing something here?

dbzTHEdinosauer · Posted: Fri Jul 30, 2010 3:06 am

zoned decimal definitons:

Robert Sample · Posted: Fri Jul 30, 2010 4:38 am

COBOL won't directly display a COMP-3 or COMP value. They are converted to USAGE DISPLAY as part of the DISPLAY statement. However, in doing so COBOL does not do anything with the sign, so X'97994D' is converted to 'F9F7F9F9D4' when displayed.

dbzTHEdinosauer · Posted: Fri Jul 30, 2010 5:11 am

thx for the post Robert.
I thought about the effect of posting my little chart,
without explanatory comments;
it would just confuse the TS.
your comment has avoided that.

saurabh39 · Posted: Fri Jul 30, 2010 12:51 pm

@Roberts - Thanks for the comment...Now I am clear about it.

saurabh39 · Posted: Thu Aug 05, 2010 12:47 am

One more question..........1234 will be represented as 4D2...which is fine...but -1234 is represented as FB2E. I checked one of the old topic...and dick replied zero minus 4D2 will give FB2E...which indeed is correct......but i did not get why i need to subtract with zero...my understanding was negative number will be stored as two complement.....so why should it 4D2..be subtracted from zero instead of binary equivalent of 4D2 being subracted from its two complement.

Robert Sample · Posted: Thu Aug 05, 2010 1:28 am

Dick gave a shortcut method. Technically, what you do is flip each bit of the value (zero becomes one, one becomes zero) so the zero becomes F, 4 (0100 in binary) becomes B (1011 in binary), D (1101 in binary) becomes 2 (0010 in binary) and 1 (0010) becomes D (1101). However, you then add 1 to the value so the D becomes E (1110) (again, for technical reasons beyond the scope of a post). Hence 04D2 becomes FB2E.

saurabh39 · Posted: Thu Aug 05, 2010 11:29 am

Thanks again Robert. The mistake I was doing - I was trying to find out complement of 4D2, instead of 04D2 as a result I was not getting intended result.

Just to rephrase, you first found out one complement of a number and then added one, which is another way of finding two's complement.

But your quote -

Robert Sample · Posted: Thu Aug 05, 2010 4:39 pm

Basically, the negative values should be ones' complement; however, since two's complement allows addition and subtraction to proceed without requiring looking at the signs of the numbers, IBM used it for their negatives. Also, one's complement allows for two representations of the value zero while two's complement does not -- which can cause problems as you can imagine. But you do have to remember to add that 1 after flipping all the bits.

I didn't intend to withhold information -- but we don't need to discuss AND gates to talk about how COBOL IF logic works. There's a lot of background to most of these questions that is noise (that is, adds to the length of the answer without increasing the clarity of the answer). Anytime you want more of the background, though, I'll provide as much as I can.