Re: How to prevent CPF5026

On Wed, 2021-06-09 at 10:23 -0400, dfreinkel@xxxxxxxxxxxxxxxxx wrote:

I have a program that writes records to a file, nothing new. However,
occasionally there is a situation where the record already exists and
I need
to add the quantity to the total quantity field. That is nothing new.

 

I do a WRITE(E) followed by IF %ERROR() and then do a update to the
record.

 

This works well.

 

However, when the system detects the duplicate record, it sends out a
CPF5026 to the job log with an automatic reply C.

 

I do not want to chain first and then do a WRITE or an UPDATE as
needed as
99% of the records will not exist.

 

How can I prevent the CPF5026 from being generated?

 

I don't think you can stop it, however (I'm going from memory here, so
wet finger in the air guess) you could use QMHRMVPM to remove it. You
may need to also look at retrieving it first QMHRCVPM. Just depends
where the message is within the call stack. Again from memory (wet
finger again), I did something similar to retrieve/remove each type of
message when a write/update failed (although possibly it was an RPG
RNXxxxx message and/or subsequent diag message) so I could replace it
with a much more user friendly message for message file display output.
Something along the lines of replacing the standard duplicate key
message with a "could not update bibliographic file record due to
having a duplicate key" with the second level text containing
information from the previous rpg/runtime/diag message.

Normally I'd leave the standard errors alone if they are "in the job
log" and not actually presentable to the user (as they are on a message
subfile)... although I could see this being an issue if there are a lot
of errors/hits. My thinking has always been: what the user cant see
makes no difference but if they can see it then replace it with
something less cryptic.

 

Darryl Freinkel

A4G

Telephone: 770.321.8562 Mobile: 678.355.8562