DSXchange

Problem is the Usage Analysis data is easily broken so things don't link up properly. Also, if people don't pull the metadata from the repository and just type it in for instance, there is no linkage. Something to keep in mind. I use Chuck's job to farm this information from an export and then I loa...

Two columns and Column Generated sql. First one is marked as a Key column and its derivation is "1". Second one is a non-key column with seqname.nextval as the derivation. Put in dual as the table name. Do not add anything in the where clause. You'll end up with something like this: SELECT 1, my_seq...

Where's Chuck when you need him?

To quote Mr Smith:

Try my List all files and tables used by jobs in a dsx file tool.

You will find it on the DataStage Tools page of www.anotheritco.com

In other words, just type something in.

It shouldn't be an issue.

Is there a way to not have it think it's transactional work? For example, in the OCI stages you can change the Transaction Isolation property from 'Read Commited' to 'Read-Only' - is there an equivalent in whatever stage you are using?

Comes with the product... document is called something along the lines of the Installation and Upgrade Guide oddly enough. Make sure you order the 7.5 media and read the one from that version. In a nutshell, the process should be as simple as: 1. Export all of your current jobs 2. Install 7.5 on the...

Do you mean this error appear because the SAP pack used is not paid ? No... Roy is just trying to find out who might best be able to help you. Those kind of errors usually mean the stage hasn't been configured correctly. Can't help specifically on SAP as I haven't used it, but I've seen it before u...

You are writing records with null key fields to a hash file, and that is a Bad Thing.

It takes no time at all to write empty files, so 'quicker' to just append? I don't think so. Better? Perhaps... that's up to you and your job designs.

If you choose to append, make sure you understand that only one process can be writing to a sequential file at any given time!

First suggestion would be to not use User Defined SQL for things like this, it's hardly ever necessary as the derivation field (and switching to Column Generated) can handle sql like this. If you insist on rolling your own, at least try the other first so that you can see what the stage generates - ...

If you stick with the 'Timestamp as a String' approach, don't forget the seconds!

Opening and closing the Sequential File stage creates the file. Writing no rows to it in the job means it will be empty when the job finishes. If that's really a concern, you could write something to delete them 'after job' if they are empty.

Yes and no. FMT is a Server function and Server functions / routines are available in a PX job via the BASIC Transformer. However, there are a number of caveats to using the BASIC Transformer in a parallel job, so your best bet is to learn the PX equivalents and stick with those.

The Num() function checks to see if something is numeric or not... interesting that the answer it returns for you is 0, aka false. In BASIC, doing math on a string does an implicit conversion so the typical trick is to multiply the field by 1. Not sure how well this all plays out in a Parallel job.

Why not just write a sql script to do this? Not every nail has to be whacked with the DataStage job hammer.