DSXchange

Try using NONE. That may work except for a few characters at the high end of the range (single byte values 248 through 255). If that works, try some of the ISO 8859 variations - there are twelve subsets - but I don't have the documentation with me. You may also need to set some database-specific env...

Looks like you've asserted that the input is sorted, and the Aggregator_2 stage has found a row not to be in the indicated sorted order.

DSguru2B wrote:Really, where....

Just before "Last edited by DSguru2B on Fri Dec 22, 2006 7:50 am; edited 1 time in total"

What are you trying to accomplish?
The join between DS_JOBS and DS_JOBOBJECTS is made as

+++ Error: missing apostrophe. Out of cheese error. Redo from start. +++

Beware that you must assign a value irrespective of the test. So you need to decide values for variables a and b when x <> 1, and for c when x = 1. Sometimes you need to leave the value unchanged; in this case assign the stage variable to itself. For example, consider a stage variable svCounter that...

Do you need to find it? Trimming leading space characters works just as effectively if there are some and if there are none. You could always compare the trimmed version with the original, perhaps in stage variables, then use the trimmed version as you wish.

Perhaps it needs a trailing semi-colon in the OPEN command?

It's "hashed" file, not hash file. 1) You can create a server shared container and use that in a parallel job, provided that the restrictions described in Chapter 2 of the Parallel Job Developer's Guide about incorporating server components are observed. 2) Yes, but the job that dumps the hashed fil...

... which you have duly reported via your support provider?

As I commented on Vincent's blog, the Modify stage is primitive. Deliberately. That's the secret of its success - it's fast, because YOU do the work.

An orchadmin command is available on both platforms.

Does the BuildLists job append or overwrite? Is "Allow duplicates" set?

Entire in an SMP environment should share the records in the (virtual) Data Set across all nodes. There's nothing documented that suggests the behaviour you describe.

Not possible. The range map is binary, and in an undocumented proprietary format. Look in the job log of the job that writes the range map; it may give some clues. Otherwise, review the design of that job, and verify your ranges from that. When using the range map to perform range partitioning, add ...

It's an even better answer if you've been bothered to document your job design.

Axiom: there is never time to retrofit documentation.
Moral: do documentation first.