DSXchange

It makes no difference when reading. The data themselves are not padded. Declare it as VarChar(6) or make sure that the data field always contains precisely six characters.

It will never be a better solution.

Open the Cleanup Resources tool in Director. On both windows (processes and locks) choose the Show All option. Can you see the job name in the Locks window? Get the process ID or user ID associated with the lock and find that user in the Processes window. This will be the user who owns the lock on t...

Every time you use a CPU cycle you "degrade performance" for some particular meaning of "performance". On the other hand, "degrade performance" is a meaningless term if you don't have a baseline measure against which the putative degradation can be compared. Writing data to disk unnecessarily is one...

So the problem has occurred trying to open one or both of the files mentioned in the Merge stage. Check their pathnames carefully, and that the executing user has appropriate permission not just to the files but also to all the directories in the pathnames to the files.

"LEX" is not a valid value for a Char(6) field, because it does not have six characters.

Where's the grouping occurring? You're probably looking for a sort key change, not a cluster key change.

My bad. Poor attention to detail there. Doing four other things at the same time.

Reset the job in Director. Find the event "From previous run...". Tell us what's in that event.

Welcome aboard.

Usually when this occurs the compilation report generated at the end contains additional information (for example which particular variable was not defined). Can you check that report to see whether there is any further diagnostic information?

It might also be a timing issue. After all, what you look up from is not the Oracle table itself, but the virtual Data Set associated with the link. Therefore, if the LUT_CreateOp operator had finished before you truncated the table in Oracle, all of the reference data would still be present in your...

One installs with DataStage. It is in a directory called Configurations and is called default.apt - beyond that, you need to construct others to your requirements specifying pathnames and other resources on your file systems. Coverage in online help and in Manager Guide (because this is where the Co...

One of the basic arts to being a DataStage developer of any merit is the ability to glean all the information there is to be had from the error messages that are generated.

Yes, tediously, through Transformer stage stage variables.

Why not specify "Don't sort, previously grouped" in the Sort stage - you can still generate the cluster key change column, at virtually no cost - what you would need to do in a Transformer stage is built-in to the Sort stage.

You have provided no information as to the methods you have used to effect your ETL. How, then, could it be reasonable for you to expect us to improve what you have done?