DSXchange

Have you thought that RT_BP676.O may not exist because you do not have permission to create it?

Are you in the group associated with the project directory? If not, the permissions string you posted indicate that you can not change the contents of the project directory.

How is the ODBC connection for your MetaStage directory configured? Do you have Import Categories already defined?

A "hash file" is used for shaping a block of hash. The term you need is "hashed file". If you've been diligent with your metadata management, making sure that hashed file table definitions are properly stored in the Repository, then you can use Usage Analysis in the Manager client to answer the ques...

In older versions (earlier than 6.0?) the default was 255. Nowadays it's 768, as Kim observes. Larger keys have implications for sizing the groups (pages) in the hashed file; after all, the entire key must be able to be stored. For example, the combination of 768-character keys and 512-byte pages si...

Another way to think of it is that the "file dictionary" is a separate hashed file in which the metadata for a hashed file is stored. Every hashed file is created with two portions; the data portion, which stores data only, and the dictionary portion, which stores metadata. The keyword DICT can be u...

Are you REALLY running DataStage version 4 with Oracle 9 plug-in stage? I do not know that this combination is supported.

The Match stage does not have this functionality. Ask yourself how you would do it with a pair of database tables? The join just gives you the matched rows; the others do not appear unless you create a second query. Then figure out how to do it with DataStage. Merge stage is not the solution, since ...

I doubt it. I think you need a different function, such as Substrings(). But that's just a guess, since you will not reveal what you want to accomplish.

What stage type are you trying to use?

Almost certainly a permissions problem. You need to have write permission (to store the message handler definition) but appear to lack it.

Make sure that the data are appropriately partitioned, so that all values in each group occur in the same partition. Otherwise you will get broken sequences.

There is no limit.

But the job still needs to be maintainable, and not of such complexity that diagnosis on the basis of warning messages becomes incomprehensible.

In my experience, it is better to have more simple, modular jobs than a few complex jobs.

Your message handler problems are almost certainly permissions related.

Access violation (also known as SIGBUS) results from an attempt to access a memory address that your process does not own. This might be that your memory space has been exhausted, or it might be an attempt to stuff 15 characters into a Char(12) data type, for example. What is your rows/transaction s...

In DataStage Manager, choose the Export menu, then choose "to MetaStage" from that menu.