DSXchange

Hi,

we have a working job which selects data from DB2 and sends it through a XML-Input stage.
This jobs works fine in DS 7.5.3 but after migrating to DS 8.1. we get problems with the german special characters which are no longer readable.
So we got some kind of codepage problem.
Is there anypone who has experienced the same behaviour or does anyone know what has changed in the XML Input stage between the DS versions?

Some tests we aready did:

• * We are able to extract the data from DB2 and write it to a file (wwithout XML stage) without problems.

• * If we do the same wit an additional XML Input stage the data in the file is goofed up

• * If the codepage for the job is changed to UTF8 and the data is read as NVARCHAR (which in my eyes is strange for a DB2 database) from DB2 the data in the file looks fine when doing a view from the stage but it is goofed up when you look into the file from the OS

Thanks in advance
kind regards
Michael

Where exactly are you with 8.1 ? While I haven't heard of any NLS issues in particular, there were many other things wrong with the initial xmlPack in 8.0, and a few subtle things that remained in 8.1 ....a first check would be to see if you have you applied the overall FP1 for 8.1 ?

It would be the overall fix pack that came out about a year ago....it should have most of the cumulative corrections.

This is not advice yet for you to go install that pack, but just to get specifics first on your environment. It may be worthy to take this one to whoever your support provider is, once you know the additional details.

Ernie

eostic wrote:Where exactly are you with 8.1 ? Ernie

Hi Ernie

we are on 8.1.0.1
so the Service Pack is already applied.

In my eyes thing that worked in th eold version should do in the new one without any changes unless documented otherwise.

Best regards
Michael

Wish I knew more about the NLS guts for you. The xml stages went thru a fairly extensive overhaul from 7.x to 8.x....mostly to add/replace the generally open source apache parser and processor that were used with Ascential with the ones required from an IBM perspective.

I have no idea of that has anything to do with the problem you are having, but it wouldn't surprise me.

It's equally possible that there is a code page tweak that can be made somewhere, but if there is, I'm not aware of it.

Ernie