DSXchange

Because a Lookup stage generates what is called a "composite operator" made up from the two operators LUT_CreateOp and LUT_ProcessOp. These need an intermediate (virtual) data set through which they can pass data.

Given the age of the previous posts, this might have been better as a new question. However, the simple answer is that the solver made sure that the host name of the services tier machine (and, presumably, engine tier) are correctly recorded in all pertinent hosts file - in short, that each machine ...

Yes. Create explicit triggers for Warning and Failure. These take precedence.

No. And you have to install Information Server in its entirety, not just DataStage.

IBM do have a migration tool to help with migration to version 8.5 (which may - I do not know - be able to do version 8.1 migrations). This is only available form IBM via their "concierge service".

No, you are almost certain to get incorrect results. And performance is likely to be woeful or you'll get "row out of order" fatal errors.

The Switch stage is closest to the switch construct in the C programming languages. It is quite unlike anything in SQL. The closest SQL construct is CASE, but even CASE allows for expressions - switch does not, being driven purely by values.

The Filter stage uses SQL-like WHERE clauses.

By inspection alone you can not answer this question. You need the metadata. This should be able to be obtained from whoever created the file.

The "6'0'R" formatting construct only works in BASIC expressions - only works in server jobs or BASIC Transformer stage.

In parallel jobs prefer

Create a new job (don't use Create Copy) and copy/paste stages to the new job.

That's too simplistic. You not only have to use a key-based algorithm, but you must also make sure that the partitioning is based on (at least the first) join key and that the algorithm will yield sufficient distinct values to spread data across all the available nodes.

Do you have a particular need to have RCP enabled?

Nice to know the speed of your memory, but it's much more important to know the capacity of your memory. You can use Link Partitioner and Link Collector stages to split your data stream into parallel streams that can be processed in separate processes, using (mandatorily) inter-process row buffering...

People, it's a hashed file, not a hash file.

A hash file is a tool for shaping a block of hash.

Even without one.

In terms of the original poster's final question, you can control a "grouped transaction" in an ODBC stage. As to the rest, please provide an English description of what you are trying to achieve.