DSXchange

Hi at all

I'm a new entry in the forum and I'm new with using datastage.

I want to understand, if possible, how to datastage pass records (or rows) between stages when job is running. Suppose we have the following situation in Server Job:

seqFile1 ----> trn ----> seqFile2

when job is running how to pass record between seqFile1 and trn and between trn and seqFile2?

How to records, without setting In-Process or Inter-Process, pass between stages? .... all together, row by row or in other way?

Thanks in advance and sorry for my basic english

If there is no IPC stage between a passive and active stage, it process the data row by row.

Yup. As noted without any buffering involved each row is processed from source to target, one after another.

Welcome to the forum and to DataStage....

As far as passing "records" or "rows" (the definition of which can be a bit ambiguous, but for the moment we'll assume them to be the same --- a single "tuple" of relational style, with one "row" of one or more "columns") between Stages, that's basically the normal behavior of DataStage....once you start writing Jobs and running them, this will become more obvious.

Ernie

Thanks for the immediate responses.

Now it's all more clear. But I have another questions to ask to you.

If a stage, in server job, process one row at a time, why with a buffer the performance improve?

Also with the buffer stage process row by row.

With a buffer more than one row can be passed between stages. Think of it as a producer-consumer model - the producer stage is writing rows into the buffer as fast as it can; the consumer stage is reading from the buffer as fast as IT can. The buffer mechanism can resist input if it becomes too full.

thanks, you have been very clear.