LastRowInGroup corrupts non-key fields

ecote · Post by **ecote** » Tue Jun 21, 2011 9:08 am

When using the LastRowInGroup function, the data in the non-key fields get corrupted with control characters.

At first I thought it was an input data quality issue but I was even able to reproduce the problem with a RowGenerator stage.

The job I used to reproduce the problem is very simple:
RowGenerator (2 key fields, 1 non-key field) -> Transformer (with a sort on the 2 key fields; running in sequential mode like all other stages) -> a sequential file stage.

The function is called on the primary key field (first field).

I only get corruption after ~200,000 groups and no corruption is found when I remove the function.

Is anybody aware of this problem? I opened a PMR with IBM and will post the solution if found...

Thanks

IBM Analytics Champion 2009 - 2020 · Post by **asorrell** » Tue Jun 21, 2011 12:47 pm

No - I haven't seen that - but what release are you on exactly?

ecote · Post by **ecote** » Tue Jun 21, 2011 2:00 pm

asorrell wrote:No - I haven't seen that - but what release are you on exactly?

On version 8.5

ray.wurlod · Post by **ray.wurlod** » Tue Jun 21, 2011 4:53 pm

Have you installed Fix Pack 1? There are lots of things fixed in this. (You can read what they are at Information Center, you can download FP1 from Fix Central.)

ecote · Post by **ecote** » Tue Jun 21, 2011 6:44 pm

I'm aware of FP1 but did not find anything related to this in the release notes. I'll wait for IBM's response and post here.

ecote · Post by **ecote** » Thu Aug 04, 2011 12:10 pm

IBM confirmed this is a bug... Waiting for solution. I will post it once received

ecote · Post by **ecote** » Tue Aug 30, 2011 7:35 am

For those interested, IBM's patch is JR39963 (currently not available for 8.5.1 but just for 8.5). The patch for 8.5.1 will be available later this year.