Archives of the TeradataForum
Message Posted: Tue, 28 Aug 2007 @ 23:41:21 GMT
It sounds like you have a lot of moving parts to examine. However, as you imply the tpump and MQ pieces are pretty robust.
On the other hand, it is difficult to make any statement on the quality of the cobol program and the 'c' program.
In my experience there are three types of people:
* Those that know 'c'
* Those that say they know 'c' and
* Those that think you collect shells and make sandcastles at the 'c' side.
It is entirely possible that the new columns you are accessing have never been properly tested in the 'c' program. If any of the columns are char or varchar, it is entirely possible that they are getting corrupted due to any number of common pointer and buffer problems.
So the 'c' program is where I would start - if it were me.
Obviously you can narrow this down by doing a couple of things.
1) Sample some records from the MQ and see that they have the correct data in the correct places.
2) Sample some records from the output of the 'c' program (is that also an MQ?) and see that the values have been corrupted.
3) Examine the parameters passed to the macro by turning on logging and see that the values are corrupted.
Another common problem is that the field layout in the tpump job differs in some fashion to the records output by the 'c' program. I've experienced this a number of times - even where the layout was agreed in advance, for whatever reason the "other side" of the record layout decided he/she would use a different layout/format. In one case, the layout changed every week - content was the same, just the layout and formats changed, but that is a whole 'nother story.
Good luck with it
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|