LET GO OF YOUR TECHNICAL BACKGROUND.
If you are a master of PL/SQL, forget it.
If you are a master of C/C++, forget it.
Approach DataStage with a new insight -- the BUSINESS BACKGROUND.
"I need to ensure that all customers' sales are collected together."
Input -> Aggregator -> Output.
Simple! (Life is not simple though...
![Mr. Green :mrgreen:](./images/smilies/icon_mrgreen.gif)
* * *
Trust me on this -- it takes SO MUCH less time to do the coding for DataStage. Revising based on new rules are so much easier. Unfortunately, dynamic fields are not fully realized (Column Propagation on the GUI is a nightmare for PX).
Once you get a good familiarity with DataStage, you can slowly reintroduce the other tools for other issues. For example, transactional control? C/C++ or PL/SQL is much better than DataStage for that, so for ODS databases, it's a recommended step. However, you do NOT need transactional control on EVERYTHING so much that DataStage is unuseable. There are many logic solutions to address that issue within DataStage.
Just let go of your PL/SQL background, and start from scratch.
-T.J.