DSXchange

Yes. What you're missing is "All the columns". A lookup on a hashed file can only use the key, the whole key, and nothing but the key. Partial lookups are not supported by READ. Partial lookups are supported by the query language but, without any indexing, each of those will require a full...

How do you or your client know that there are "performance problems"? What is "performance" in an ETL context? Maybe the issue is more one of expectation management.

Search DSXchange for some small fixed part of the error message; you will find that your question has been asked and answered before.

Good time to learn. Look at the examples that ship with the product, and get a beginners book.

No, it doesn't mean that at all.

*Topic Relocated - Content Editor* In version 8 server jobs execute in a "cocoon" that has to interface with services on the parallel execution architecture. That can make the server job, unchanged, take longer to run than on an earlier version. Even so, note Ernie's point about server job...

Or create a DataStage job to do it. Prefer a server job because these have inbuilt stage types for reading from hashed files.

It may be that there are no locks.

Open Administrator client Command window and execute LIST.READU command. This displays the contents of the local repository lock table.

In Director go to the View menu and disable display of categories/folders. The count of job names will be in the status bar, and the display will contain a list of all jobs, which you can sort based on any of the column headings. You can then use Director's report printing facilities if you wish; th...

DSD.BCIPut only invokes OCONV for date and time data types, so I'd start by looking there.

And DO do something about naming stages and links meaningfully.

Question: why isn't your client (or you) hiring someone who knows what they're doing?

Then you have deprived the owners of those data of some of their data.

On whose authority were those characters declared to be "junk"? Did you get agreement from the business to delete them? Who, in the organization, is the data steward with responsibility for this item of data?

There is also a Surrogate Key Generator, which is what I believe nagarjuna meant. But a Column Generator would do as well; either way you will need to initialize the sequence. And beware of parallel processing.

Server job or parallel job?

It's probably bearable particularly if you have a short auto-purge interval. Otherwise create your own custom routine or job control code.