Page 1 of 1

Sequential file behaviour

Posted: Wed Sep 09, 2015 2:07 pm
by soumya5891
I have job where it reads from sequential file and stores the data in dataset.There is one column in the sequential file named VERSION(say) and it's defined as VARCHAR(1).

Now in the file if any records are coming say 'AB' for VERSION then it truncates up to 'A' in both 8.5 and 11.3.But in 11.3 am getting one warning like the below where in 8.5 no warnings have been generated.

External string too long. Imported only 1 out of 2 external characters. Ignoring last 1 characters up to delimiter.

I have checked the message handler also,bit nothing is defined in 8.5.

Any idea on this?

Posted: Wed Sep 09, 2015 2:59 pm
by chulett
Working properly now, wasn't before.

Posted: Wed Sep 09, 2015 3:41 pm
by soumya5891
Yes,from my understanding also the current behavior is supposed to be right.But in current prod(8.5) the data is truncated in that way and that field has no further significance, but in 11.3 it's throwing warning and the job aborts

Posted: Wed Sep 09, 2015 4:19 pm
by ray.wurlod
You seem to be relying on something that was not properly checked in version 8.5 but now is being properly checked.

Posted: Thu Sep 10, 2015 8:23 am
by qt_ky
Agreed. I have seen numerous jobs give new warnings in 11.3 as compared to 8.7 (no warnings), and in all cases so far, the job has had something coded incorrectly, like a data type, or a delimiter, or other setting mismatch. Fix the job design in 11.3 and the warning disappears.