DSXchange

There will be reclaimable space, but it will not be reclaimed - it can be re-used, as it is marked as "free", but the physical size of a hashed file never reduces until and unless a RESIZE operation is performed.

Stop trying to aggregate (group by) a column that might contain null. You've been lucky for two months, now your luck's turned bad - there's now a null in one of the grouping or aggregating columns in the data.

No. An OLE server is a way to expose ActiveX objects - it's a Microsoft invention, not special to DataStage. DataStage Objects, whether for ActiveX or for Java, exposes 13 different objects that you can read about in the UniObjects manual (downloadable from IBM UniVerse manuals web page). OLE plays ...

The project may well be intact. It may be the system tables that refer to it that need verification. Search the forum; you are not the first to have encountered this particular symptom.

On the columns grid of the input link, Save.
On the columns grid of the output link, Load from where you saved.

1 and 2 When DataStage first appeared at the beginning of 1997 it ran on UniVerse as its engine. When IBM acquired UniVerse in 2001 the two products needed separate names. Hence "DataStage Engine" was born (as a name). 3 RT_BPnnn.O where nnn is the job number. Note, however, that this is not execut...

Have you tried, using a regular Sequential File stage?

The "intermediate table" approach could be gotten to work in a server job, but is expressly prohibited - as a "blocking operation" - in a parallel job. Unless you're prepared to re-design using two jobs, this suggested approach is inviable for you. MetaStage ceases to exist in version 8, so it's pro...

Yes, by detecting the failure condition, sending a row to a separate output link and, in the constraints grid, setting that link to fail the job when its row count reaches 1. I advocate always creating jobs that do not abort, as that way you can maintain control more completely. In that case a diffe...

Some administrators restrict access to chmod, chgrp, chown, etc., just as a matter of course.

"DataStage Objects" is a marketing term, it is DataStage's equivalent of UniVerse's "UniObjects" (which, prior to VMARK's merger with UniData, was called "UniVerse Objects"). Each is a proprietary way of exposing objects on the client that connect fairly directly with the DataStage/UniVerse database...

theseals wrote:I have already made it sufficiently clear to the development team that the practice of changing column definitions within a job design should be avoided if at all possible.

Go with that. But stronger.

theseals wrote:I have already made it sufficiently clear to the development team that the practice of changing column definitions within a job design should be avoided if at all possible.

Go with that. But stronger.

Simply, "permanent storage" means the data files of Data Sets, File Sets and Lookup File Sets. "Temporary storage" covers files created during (and removed after) certain processing, such as sorting. Any operator that needs more memory than it has been allocated will create temporary files on the sc...

Information about the production environment is specific to site. You need to find out, from whomever administers the systems upon which you are working, the relevant hostnames, user IDs and passwords - you have, of course, made all of these job parameters. The production project may be a protected ...