DSXchange

Search reveals it all. Thisfor instance.

Is this server job or px job?

Make sure you have identical metadata both, while building the hashed file and while referencing it.
Also make sure that you trim the key fields while loading the data and also during lookup, trim the source keys.

Great. You can mark your post as "Resolved" then.

You have to do inserts and updates seperately. Not in one query.

If your referring to the bulk load stage then both would be the same I suspect. I dont have any test results to back it up. But technically both use DB2's bulk utility. Hence it should be the same. If your talking about the DB2 API stage and DB2 enterprise stage then the latter is way way faster as ...

You dont have online manuals installed on your desktop? It gets shipped along with the product. Look for it under the Ascential directory in your local drive. Anywho, guess you have to specify keys from both the files in your output as per the following from the manual: You must include a First Fil...

Do you see any mention of g++ in the old environment's dsenv file? See what variable its path is assigned to. Set that variable with the new environment compiler's path.

What does the manual say about Merge stage?

Can you double check at the database level, wheather this field really is nullable or not.

Now the OP cannot delete the post can he? If you reply to a post, the OP cannot delete it anymore. Only the moderator can do it from there on.

I doubt if the same function is present anywhere else. Use the transformer stage.

I havent worked with the Merge stage but once the file is created, you can read it as it is and not propogate the extra column through the transformer.

Query the database with an orderby on the key.

Well some path is not set in the dsenv then. g++ cannot just be accessed just like that. If it worked in your previous env. that means that the path to its bin directory was set in the dsenv file. Your admin should know.

O ok. Now that i re-read the question I see. The OP means call a transform routine from a before/after job subroutine. Right???