DSXchange

If your rows average 98 bytes, is that like saying the smallest row is 1 byte and the largest is 200? Ray is suggesting maybe you have a lot of rows that are way above 98 bytes. You could be bumping against the 2.2 GB limit on a 32bit hashed file if your write cache is holding a lot of data and can'...

Chances are a NULL value is making it into a calculation. This causes ABNORMAL TERMINATIONs of the job.

If a job has a message like this:

Program "JOB.1246365973.DT.1405220768.TRANS1

it's a DS Server Transformer stage, not a PX stage.

For some reason, I was thinking the Folder stage was Windoze only. I checked the 7.5 documentation and it looks like it might work. I'm pretty sure that it's one-row to one-file, unless you pack all rows for a file together. The only way to do this is to SORT by the filename column I suggested, then...

Welcome to our forum. Your requirement is not one that DataStage is meant to solve. Of course, you could use job parameters for the file names, but because you don't know the file names ahead of time you can't use job parameters. What I would suggest is that you add a column to the output file, say ...

Are you creating the file or is it coming to you? If you're creating it, where are you getting it? You'll have to put some derivation logic on the interested columns.

If the file is coming to you this way, you'll have to address the file prior to reading it with PX.

How big is the file, and what do you want to do about the low ASCII values? If the file is small and you want the low values swapped to a space, a simple filter using sed could swap out the values. Anything larger and complicated could require some work.

Manually recreate the RT_LOGnnn directories and contents, or write a script to do it.

Can you paste the mkdbfile command logged in the job log? I'm curious to why the hashed file is not growing properly. It "feels" like you played with the Split/Merge settings.

You should never do this again. Internal repository files should NEVER been messed with. Your project is basically corrupted at this point. If you left the RT_LOGnnn directories behind then the easiest solution is to go to an existing job in Director and clear its log. Then, go to the directory that...

Are you sure you created this hashed file? The OVER.30 file is larger than the DATA.30 file. My guess is that you need to remove this hashed file first and try again. You'll need to do it anyway because of the corruption. Make sure you have the disk space on the file system, but watch this hashed fi...

We've written our own job control that takes a table of jobs and their dependencies and manages the execution. Part of that include throttling, but when you're talking about PX you've lost control over resources. Node pools are about the only way to deal with resource allocation, but even that's tri...

Renaming a file in DOS is "rename oldfile newfile". As for the best way to loop is either custom job control with a For Next loop or a Sequencer with a looping structure.

Parallel datasets are stored within the PX workspaces. You must use PX commands to browse or list those datasets. If you need a Server job to access PX datasets, you must pull out the datasets and place them where the Server job can see them.

You'll have to write a parsing function. Unless someone out there has already written one that is PX ready.

That pesky transformer stage allows you to have a reject link.