DSXchange

Get into the habit of renaming stages and links with meaningful names as soon as you place them in the design area, and renaming sequential file columns to meaningful names while on the Define tab of the Import wizard.

DSLink9.Field003 is anathema.

Please explain the circumstance for which you require this functionality in a server job. Server jobs, as a general rule, do not make much use of environment variables. It can be done, but there is no out-of-the-box routine for doing it. But without good reason I am not going to waste time creating ...

What's the largest number that DataStage can process? 38 digits - this is documented in the DataStage BASIC manual and other places. How many digits in 100! ? You CAN do it in DataStage - but why? To do it in DataStage you will need to use "string math" functions (SMUL() in paricular), which work wi...

In production this function will not be optimally efficient, since it needs to go through the CASE construct for every row. A much improved version will appear shortly - I'll advise on this thread, and document why the new version is, to quote Craig H, "more better".

pravat_dw wrote:Can any one help me out.

We'd love to help you out. Which way did you come in?
(-- Spike Milligan)

Any reason not to post a question that is obviously about a parallel job in the forum for parallel jobs?

Any reason not to post a question that is obviously about a server job in the forum for server jobs? There's no such thing as a hash file (other than a tool for shaping a block of hash). How did you attempt to create one? Assuming the thing that you attempted to create was a hashed file, how did you...

When you run the job from Director, select the Tracing tab in the Job Run Options dialog and enable statistics collection for each active stage. The gathered statistics can be used to provide finer detail as to where each stage is spending the majority of its time.

PX moves data in chunks of at least 32KB. If you have less data than that - or even less than 128KB in my experience, you're not really going to get good parallelism. (For fixed width data the transport block size is even larger; 1MB by default.) Try your tests with large data volumes - at least 10MB.

If you can have both "" and NULL, I would use a non-default representation of null, such as "<NULL>" - something that will never occur in data.

What are the data types of the columns mentioned in the WHERE clause?

Check the first entry for the job in the job log, to determine the values that were passed for the parameters.

Any mechanism external to DataStage will need a sequential process to generate unique keys. The only way truly to use fully-scalable parallelism is to initialize a lookup data set with the next available key, and then use a formula based on the partition number and partition count to add to that. Ex...

View Data will get rows from the first partition first. That's why you're seeing only odd-numbered rows. If you were to scroll down far enough (and select enough rows to read) you will find the even-numbered rows.

Stage variable svRowCount. Initialize to 0. Derive as something like