S0C9 Error fixed-point divide exception

saji123 · New User Joined: 15 Jul 2009 Posts: 19 Location: Chennai

Hi
I am getting Abend S0C9 while submitting the JCl error is shown as
The system detected a fixed-point divide exception (System Completion
Code=0C9).

But I havn't used and Division in my Code. Can anyone tell mw what other reason will be there for this error..

Ketan Varhade · Posted: Mon Sep 07, 2009 2:06 pm

Hi,
while trying to edit or browse a PS dataset with LRECL 0 the File Manager abends with S0C9 check the LREC of the file

saji123 · New User Joined: 15 Jul 2009 Posts: 19 Location: Chennai

LRECL is also fine. I am moving a COMP value to non-COMP value and doing calculation on this variable. Will it cause an error. i have declared both the COMP value and the non-COMP value with the same size.

Ketan Varhade · Posted: Mon Sep 07, 2009 4:02 pm

It won't give you a SOC9 for this reason.

Robert Sample · Posted: Mon Sep 07, 2009 6:52 pm

Terry Heinze · Posted: Mon Sep 07, 2009 7:09 pm

saji123,
It would help if you posted (copy/paste from your program) the field definitions and all relevant code and error messages involved.

mmwife · Super Moderator Joined: 30 May 2003 Posts: 1592

Hi Saji,

It would also be nice to know what COBOL stmt caused the Abend. Have you done that?

If not, do a search of the posts here using "OFFSET" or "0C7" to find how to do it.

saji123 · New User Joined: 15 Jul 2009 Posts: 19 Location: Chennai

the piece of code where i am getting error is
MOVE W040-EQRFND-CREDITAMT (W040-EQRFND-SUB)
TO WS-EQRFND-CRAMT-TEMP

I have declared field as W040-EQRFND-CREDITAMT PIC s9(16) comp.
WS-EQRFND-CRAMT-TEMP PIC s9(16) value 0.

Terry Heinze · Posted: Mon Sep 07, 2009 7:31 pm

Please post what is requested. And copy/paste what you post directly from your program instead of keying it. Your source field in the MOVE statement is subscripted/indexed but you didn't post the definition of the subscript (if subscripted) or the value of it.

saji123 · New User Joined: 15 Jul 2009 Posts: 19 Location: Chennai

Oh Sorry !!

01 WS-EQRFND-CREDITAMT PIC S9(9)V99 VALUE 0.
01 WS-EQRF-CREDITAMT PIC S9(15)V999 VALUE 0.
01 WS-EQRFND-CRAMT-TEMP1 PIC S9(9)V99 VALUE 0.
01 WS-EQRFND-CRAMT-TEMP2 PIC S9(9)V99 VALUE 0.
01 WS-EQRFND-CRAMT-TEMP PIC S9(15)V999 VALUE 0.

01 W040-EQRFND-TABLE PIC X(870000) VALUE HIGH-VALUES.
01 FILLER REDEFINES W040-EQRFND-TABLE.
03 W040-EQRFND-ENTRIES OCCURS 10000
ASCENDING KEY
W040-EQRFND-CUSNO
INDEXED BY W040-EQRFND-SUB.
05 W040-EQRFND-CREDITAMT PIC S9(9)V99 COMP.

MOVE W040-EQRFND-CREDITAMT (W040-EQRFND-SUB)
TO WS-EQRFND-CRAMT-TEMP1

Error Message: The system detected a fixed-point divide exception (System Completion
Code=0C9).

Terry Heinze · Posted: Mon Sep 07, 2009 8:07 pm

Please post what is requested. What is the value of W040-EQRFND-SUB immediately before the MOVE statement is executed (you can DISPLAY it). And please use BBCode to make your post more readable.

Robert Sample · Posted: Mon Sep 07, 2009 8:08 pm

Where's your SET statement for W040-EQRFND-SUB? You want us to figure out an index error without telling us how you set the index?

saji123 · New User Joined: 15 Jul 2009 Posts: 19 Location: Chennai

SET W040-EQRFND-SUB TO 1

MOVE W040-EQRFND-ADVNO (W040-EQRFND-SUB)
TO WS-PREV-ADVNO
MOVE W040-EQRFND-CREDITAMT (W040-EQRFND-SUB)
TO WS-EQRFND-CRAMT-TEMP1
SET W040-EQRFND-SUB UP BY 1
MOVE W040-EQRFND-ADVNO (W040-EQRFND-SUB)
TO WS-CURRENT-ADVNO
PERFORM UNTIL WS-CURRENT-ADVNO NOT = WS-PREV-ADVNO
MOVE W040-EQRFND-CREDITAMT (W040-EQRFND-SUB)
TO WS-EQRFND-CRAMT-TEMP2
ADD WS-EQRFND-CRAMT-TEMP2 TO WS-EQRFND-CRAMT-TEMP1
MOVE W040-EQRFND-ADVNO (W040-EQRFND-SUB)
TO WS-PREV-ADVNO
SET W040-EQRFND-SUB UP BY 1
MOVE W040-EQRFND-ADVNO (W040-EQRFND-SUB)
TO WS-CURRENT-ADVNO
END-PERFORM

Now am getting error in the ADD statement.

Anuj Dhawan · Posted: Mon Sep 07, 2009 8:34 pm

Did you notice what Robert has pointed to in his first post? Dividing by zero is the most common cause of this ABEND. Correct the program logic error that caused the divide exception and resubmit the job.

This ABEND is caused by a fixed-point divide exception. A quotient has exceeded the register size in a DIVIDE instruction, or the result of a CONVERT TO BINARY instruction is more than 31 bits long. Don't you have any divide-by instruction in your code at all?

saji123 · New User Joined: 15 Jul 2009 Posts: 19 Location: Chennai

No i dont have any Divide by instruction in my code

Robert Sample · Posted: Mon Sep 07, 2009 8:46 pm

Recompile your code with NOOFF,LIST to get the pseudo-assembler listing. Once this is done, rerun the program and the offset for the abend should point to one of the pseudo-assembler instructions. Post the exact error message and pseudo-assembler code (5 instructions before and after the line with the error should be enough). And use the BBcode for your listing to make it easier to review.

Terry Heinze · Posted: Mon Sep 07, 2009 9:13 pm

saji123 · New User Joined: 15 Jul 2009 Posts: 19 Location: Chennai

I am pasting the code directly from the program. I got error in the ADD statement since i declared the variable as COMP. I have changed my Code a little.

My exact code is like

Robert Sample · Posted: Mon Sep 07, 2009 9:47 pm

And what do we see at 1B3E? A Divide pseudo-assembler instruction. If you got a dump from the abend, look in the TGT to find the SYSLIT address and figure out what is at offset 8. Somehow that value is becoming zero -- possibly due to table overflow.

If you haven't compiled with SSRANGE, do so and run with the run time option to turn on subscript range checking.

saji123 · New User Joined: 15 Jul 2009 Posts: 19 Location: Chennai

Hi ,
Thank you for all your support i got the solved the error.
I was problem with the input file
Thanks

Anuj Dhawan · Posted: Wed Sep 09, 2009 4:10 pm

Discussion expanded to two pages and solution is