DSXchange

Indeed it is. Any time you invoke the Run interface you get three choices - run now, create a "bundle" that the iaJob can deal with later, or invoke the Information Server scheduler for one-shot or regular schedule.

RK72 wrote:Its not about less RAM.Already 16 GB of RAM is allocated to it.We run the jobs on the server only so no client required.

Allocated in VMware?

No, I mean to ask the Notes administrator whether there are any locks in the notes database, and try to resolve why.

Add more virtual memory (lots more- at least 4GB to the server, 8GB preferred, and at least 2GB to the client) to the VMWare images. If the underlying machine does not support that, get a higher-powered (more memory) underlying machine.

Read the records as a single string, use TrimB() function to strip the trailing space characters. In a Transformer stage have as many output links as there are fields in the record, and send one field to each. Downstream of that use a Funnel stage to bring them together.

The parallel execution philosophy is "one task, one stage type". The Merge stage merges. The Transformer stage transforms.

Exactly. You can generate these constants in an upstream Column Generator or Transformer stage.

Do you always choose just to murder your jobs rather than to try to find out why they're hanging? Often the problem is a lock in a database - wouldn't it be wise to ascertain that and free the database lock so that the job can continue to completion?

Obviously the database.

Unless you can be more exact than "not working" we cannot help further.

Natural keys occur more than once each in the target database. So, unless you use complex logic (constrain also on the largest associated surrogate key, which would involve a correlated subquery), you're probably not going to get the correct result except by chance. You never did explain what "...

Full instructions accompany the software. You can also get installation information from the IBM web site. You will need to install a metadata repository (DB2, Oracle or SQL Server), a WebSphere Application Server, a set of Information Server services, at least one DataStage and QualityStage engine,...

I'm not sure what you're saying. Surely you're using something like "item number" as the lookup key, so one lookup could return buy price and another lookup could return sell price from the same source data? Of course, two copies of the reference data will be needed in memory as soon as yo...

You don't seem to have come to grips with what the configuration file does in terms of resource allocation. In particular you have used a node pool name of export, but you have left this node in the default pool anyway. That's why the same resource disk continues to be used. Node and disk pools are ...

Unfortunately, that's "the way things are". The parallel Sequential File stage is notorious for not handling embedded quote characters properly, as a Search through this forum will reveal. You can, of course, put your own quotes on and specify the Quote Character property as none, but it w...

The single argument to the key management routines is the NAME OF A SEQUENCE. The sequence is stored in the DataStage repository, however, not in the target database.