I can see that Collection, Location, Package, Package Owner and TCP Port have changed. Can you shed any light on why the new values are good and the old values are bad?
Do you mean PX or TX? You've posted in the parallel (PX) forum, but asked a question about TX (Mercator). The answer will be different in each case. In a parallel job I'd use a Transformer constrained to output only when CODE is 2. I'd store the code 1 column values in stage variables conditionally ...
Some examples of values where it succeeds and, particularly, where it fails may aid the diagnostic process! You should be able to capture these via a Peek stage connected to the Transformer stage's Otherwise/Log output.
Welcome aboard. :D The answer is no. The only version of DataStage that supports execution on a Windows platform is 7.5x2 (not the same as 7.5.2). Even this release is not supported on Windows XP Professional server; it can be installed and will work, but you're on your own if you need support. Data...
It would be better if you research this yourself. The mainframe.pdf guide is part of the standard DataStage manuals. Read about the various file handling stage types, particularly Chapters 3 and 4.
The first match "consumes" the 100 10 x row from the second Data Set. Therefore other "matches" are reported as inserts.
Try identifying both COL1 and COL2 as keys.