Hello,
I did some test with FILL-MODES for the buffers of a ProDataset. So I came to the point where I did set the FILL-MODE to REPLACE and then tried to populate the ProDataset. As long as I did not try to implement any error-handling everything was fine. But then I looked at the ERROR-STATUS and saw that ERROR was set to TRUE and ERROR-STATUS:GET-MESSAGE (1) returned “** … record not on file. (138)”.
It seems to me that the message (there is no such record in the TEMP-TABLE so I have to create a new one) gets passed out of the ProDataset FILL statement. This is the expected behavior of the FILL-MODE – REPLACE so why should I be interested in such messages? And in fact this complicates the implementation of a good error handling and makes it hard to handle the real errors.
I think that passing out the message and that the ERROR-STATUS is not reset in such cases is a bug. How do you think about it?
Regards,
Marko
Attached you can find a little procedure how to reproduce this behavior.
I logged this issue with PSC TechSupport: W008250016
ProDataSetFillInReplaceModeTest.p.zip
