CEE3209S: Fixed-point divide exception (0C9) under Language Environment

What it means

Binary divide instructions work on integers held in registers. If the divisor is zero, or the quotient is too big for the result, the processor raises a fixed-point divide exception. Under Language Environment (LE), that program check produces CEE3209S. The message says the system detected a fixed-point divide exception and gives system completion code 0C9. As with the other CEE32xx messages, the next line gives the compile unit, entry point and offset.

IBM describes the cause as a signed binary division with a zero divisor. In COBOL this comes from a DIVIDE or COMPUTE on binary fields (COMP, COMP-4, COMP-5 or BINARY), or from a conversion of a packed number too large for a binary result. In C it is integer division by zero. The same mistake on packed decimal fields gives CEE3211S instead.

IBM notes that a condition handling routine can correct the data and let the application resume. With no handler, LE ends the thread. The symbolic feedback code is CEE349.

Common causes

• A divisor came from input or a calculation and was zero for this record, and the program didn't check it first.
• A count used as a divisor, such as a number of items or days, was never set because a branch that sets it didn't run.
• A DIVIDE or COMPUTE has no ON SIZE ERROR phrase, so nothing stops the zero divide from reaching the hardware.
• A packed decimal value too large for a 32-bit or 64-bit binary result was converted to binary. The processor reports that overflow as a fixed-point divide exception too.
• A storage overlay set the divisor field to zero.

How to find the problem

1. Take the offset from the line under CEE3209S and find the statement in the compile listing.
2. Look at the divisor in that statement. The CEEDUMP lists local variables for COBOL programs compiled with the right options, or you can use a debugger or Fault Analyzer.
3. If the divisor isn't zero, check whether the statement converts a large packed value to binary.
4. Find the input record being processed when the step failed, so you can see why the divisor was zero.

How to fix it

• Test the divisor before the division and handle zero in whatever way the business rule requires.
• Add ON SIZE ERROR to COBOL DIVIDE and COMPUTE statements that can see a zero divisor.
• Give count fields a starting value, and make sure every path through the program sets them.
• Widen binary fields that receive large converted values.

What it looks like

CEE3209S The system detected a fixed-point divide exception (System Completion Code=0C9).
         From compile unit BILLCALC at entry point BILLCALC at compile unit offset +00000A10
IEF450I BILLRUN3 STEP020 - ABEND=S0C9 U0000 REASON=00000009

Last reviewed 2026-10-18 by the WND Software team. Written in our own words. For the full reference, check IBM’s documentation for your z/OS release.