If it is an Inner Join you have at least join the two tables in the DataBase, so why not sort and agg them there, that is more performant then doing it outside. If the task is to big for the database (and/or DataStage) there might be way to solve it in steps. Sometimes programming is also a kind of ...
IBM is (or may be is got) such a complicated construct. All the jokes we tell about the burocratic world are true like cycling responsiblity: You phone to somebody, (s)he says "I am not in charge, phone Mr(s) B", "B" say the same about "C" and finaly "Z" refer...
You are welcome, Arnd. I've send you a list of usefull IBM-Links for DataStage-people. Udo Neumann did a good work with this list. He presented it on our last usergroup-Meeting in Cologne.
Okay packing was since version 3.1 one of two options and no more the only one of versioning Jobs. At that time the installation was only possible by root. The Release and Package Wizzard has two advantages: (1) no need to leave the (Unix)-Server (2) no need to have any DataStage-Client-Tool for the...
Concerning telefonnumbers some remarks: (1) you should check more than "Null" perhaps if at least one number is in it. (2) There are a lot of formats which differ from region to region in the world. (3) In some countries and among countries telephone numbers have a different length (4) Bes...
If there would be an optimal "Array size" DataStage would use it. Array size is something which might make you job perform in a better time. You have to make some experiments for to find the optimal point for each Stage and Job. Concerning a caution Transaction Management it should be &quo...
we have a client-side commandline-tool called "dsmove", which can be controlled by any script like batch or else. It can: - import any ds-element - export jobs, routines and transforms (as single *.dsx-file for each element, if needed) - delete elements - compile jobs all via command line....
Dear Brah-Man, Transaction Mechanisms in DataBases was primarely made for single row operations of multiple tables. Within DataStage you can use them as described by Arnd and Asita. One Problem with large volumes is that some time the rollback segments get full and fuller. We made since many year go...
It depends how fine you want to have the timestamp. If you pass while starting your job the date/time as a parameter you are able to use it for the file naming.