DSXchange

You're only young once but there's no statute of limitations on immaturity.

To be old and wise you first have to be young and stupid.

Ask your account rep to supply, on the basis that you need to prepare and plan to upgrade.

Click on the More button to get complete output from the compiler.

Check that you have write permission everywhere in the DataStage client installation folder. Part of the pre-compilation phase takes place on the client.

Try taking a copy of the job, and compiling and running that. If that works, delete the bad one, rename the new one and recompile.

I am Loading from a Sequential File to a DB2 Table Seq File - > Transformer - > DB2 Table I need to perform a lookup on the Target DB2 Table everytime before inserting or updating records. The Lookup Should be updated everytime the target has a new or updated records. I tried having the DB2 stage a...

If Ascential\DataStage7.5.2 is where the client software is located, nearly everything in it is essential, with the possible exception of Docs (the manuals). For example, all the bitmap files are icons for the various client tools. Nevertheless, keep these manuals somewhere; you'll probably need to ...

Please delete your duplicate post
Provided that whatever stage types you are using are certified for use with Sybase 15 (check on the IBM web site) you need make no changes to your DataStage jobs.

It might be worth trying AsInteger(DecimalToDecimal()), with the first transition to a decimal number with 0 decimal places.

Your second approach requires half the memory at run time of your first approach.

Harrumph!! If you're going to use DataStage BASIC, you may as well just use the Fold() function.

But this is the parallel jobs forum...

It does tell you in which stage the error occurred. Only one stage from your design will generate an export operator in the score. Inspect the score to determine which it is (assuming you can't guess).

I don't think anyone has ever claimed that the Parallel Job Developer's Guide is 100% accurate. That's why it's so useful to have the Orchestrate manuals as well which, in my opinion, are more accurate.

Integer doesn't have a length. Data type int32 is always four bytes long. If your aggregated numbers are larger than this, you can use int64.

In a server job there is no support for right outer join other than via a Merge stage, which needs to read two text files.

You need to reverse the sense of stream and reference inputs, and perform a regular lookup. The nulls will happen automatically.