DSXchange

Active stage execution times are reported to the nearest millisecond in the job log, in "active stage finishing" messages.

You can also enable statistics collection (to the nearest microsecond) on the Tracing tab of the Job Run Options dialog when requesting a job run from Director.

At run time the architecture is different. The DataStage clients do not need to be involved. As far as database servers are concerned, the DataStage job is just another client application.

Run an after-stage subroutine to detect the existence of the bad file and set ErrorFlag to a non-zero value if it is found.

Also add a before-stage subroutine to delete the bad file if it exists.

Try preceding the command with start so that it runs in its own window. (You can learn about start in DOS HELP.)

OK on dialup. Doubt that it would be with a bigger screen. Unable to comment on screen shots - only got to see Andy's phiz. And, yes, speaking more slowly will aid those whose first language is not Texan.

Beat me to it! "MCN" is one of my all time favourites!

CLOSESEQ (or FLUSH) will create the file if any write operation (WRITESEQ, WRITEBLK or WEOFSEQ) has been performed. WRITESEQF will also create file file.

If no write operation has been performed but CREATE has been used, the file will be of zero length when closed.

The answer is a definite maybe. It really does depend on what your job is doing. If stage ---> IPC ---> stage ---> IPC ---> stage runs faster than stage ---> stage ---> stage then would stage ---> IPC ---> IPC ---> stage ---> IPC ---> IPC ---> stage run even faster? The answer is no, for the same re...

Source code is not multiply instantiated. Only the executable program is. The invocation ID ensures that each instance is uniquely identifiable. They run as completely independent process (that is, not as multiple threads in the same process).

"Create" is clever enough not to create if it already exists.

Step 3 is typically most efficiently accomplished with a hashed file that is pre-loaded with the relevant rows and columns from the target table, and kept up to date as source rows are processed.

Certainly read timeout should be non-zero. Zero means wait forever. Is anything writing into the other end of the pipe?

Technically they are different. You should use record delimiter property to record the record delimiter always. The final delimiter property is for when you might have a different delimiter following the final column, and different from the field delimiters within the record. For example 12,"Arthur"...

Reset the job in Director. Additional diagnostic information may be logged "from previous run". Ask your support provider what "code 134" means.

Under Ken's scenario you would tend to get "unable to open RT_CONFIGnnn file" and/or "T30FILE table full" errors. Are you seeing any such in the log? The disk full situation does not have to be permanent - it can be really transitory and still cause a problem. For example, are your disk and scratchd...