RE: CPF7CFD Error occurred running REXX procedure

Just thought I'd let you know that this issue has been resolved. A PMR was opened. It seems that everytime the job used an SQL statement embedded in a REXX procedure, a 'space' is attributed and the number of spaces being created was exceeding the maximum available.
The job was failing after about 5000 sql deletes had been issued.
We were finally asked if we could work around with a stored procedure. We embedded the sql in rpg instead.
Shame, it was the only REXX procedure on our system.

-----Message d'origine-----
[mailto:midrange-l-bounces@xxxxxxxxxxxx] De la part de CRPence
Objet : Re: CPF7CFD Error occurred running REXX procedure

How many programmers are required to replace a light-bulb?
None. It works on my machine. :-)

"It" is not the same thing in the working and failing scenarios.
When EXECSQL gets invoked properly and the successful effect
of establishing an SQLCODE transpires, then the expression
ABS(SQLCODE) will be processing the "absolute value of a
numeric string". When the invocation of the SQL does not
establish the SQLCODE variable\value, the effect of the
numeric scalar function ABS("SQLCODE") will fail with thee
"bad arithmetic" error.

Using the condition "If RC<>0" is not a proper means to
evaluate if the SQLCODE variable is a valid
numeric\numeric-string value. Either that conditional logic
or the later expression using SQLCODE, needs to establish
that the SQLCODE value is numeric before using the SQLCODE
value in a numeric evaluation\expression.

Note: Since the result of SQLCODE being non-numeric is an
indication of a defect in the EXECSQL [i.e. unexpected
result], I see little reason to change the code, I am just
suggesting what is wrong with the code [in the failing scenario].

Regards, Chuck

On 1/3/11 10:41 AM, Luis Rodriguez wrote:

Nevertheless, it works for me (at least on V5R3)...

On Mon, Jan 3, 2011 at 2:00 PM, CRPence wrote:

Luis, I think REXX is trying to be funny.

I typed : SAY SQLCODE

and guess what? It said SQLCODE.

How do I get it to output the value of SQLCODE?

Luis Rodriguez on 3 janvier 2011 16:31 wrote:

You are right, of course. I'm afraid I misread the parenthesis's
order.

I just ran his statements without problems. On the other

hand, it

would seem as if SQLCODE had an invalid value (?). Time

for a SAY

SQLCODE, perhaps?

That output is evidence there is no value assigned yet to the
variable SQLCODE; i.e. per no value, the string itself is output.
That is also an indication of why using the RC versus SQLCODE as a
condition to report the SQLCODE is the source of a bug in the
original REXX source statements.