DSXchange

The "trick" - which will work in server or in parallel jobs - is to create "generic" column definitions then to use user-defined SQL the whole of which is supplied through a job parameter (such as #SQLStatement#). This may, for example, need to select constants to fill the unneeded columns or column...

Even with version 7.5 I find it useful at times to refer to the Orchestrate manuals, because I've been burned in the past by errors in the Parallel Job Developer's Guide. It's good to have a cross-check available.

I'd be prepared to cut some slack here - there may be a good suggestion coming up, depending on whether the OP is locked in to using a script or not. Can always shut it down later if it wanders.

It is not different. DataStage is simply another ODBC client application that makes calls through an ODBC driver using standard ODBC API calls.

One of the pieces of advice I always give when running the DataStage Essentials class is "right click everywhere!". There just isn't enough screen real estate to make the whole of the functionality visible. For example who knew that every Director view has a filter? Or that you can view data in Desi...

UniVerse tables ARE hashed files. Ah, ignorance is bliss! Hashed files are quite able to handle millions of rows, particularly if the rows are small (key and one or two columns). If you need to go over 2GB in any one hashed file, enable 64-bit addressing for that hashed file. And the hashed file rea...

I've installed it on a two-bit XP server. Yuk!

The approach you're already taking (using xlst) is almost certainly the most efficient for larger data volumes.

The DataStage XML stages have this terrible proclivity to run out of memory.

Well spotted, Keith, and the reason that I made the comment in dark red rather than in black.

When you call it "binary 8" do you mean uint8?

If so, why can't you just load this directly with the appropriate numeric value? (I'm used to seeing -1 in an int32 column, but the details don't matter.) To me, the string_to_raw conversion seems unnecessary.

No, ma'am. We're musicians.

You're doing all the lookups in one Transformer stage? One thing you can do is to separate it out into two or three Transformer stages (or more) and use explicit IPC stages to force them to execute in separate processes. Your current design is trying to do all six lookup from one process. Three Tran...

Stage type versioning is separate from versioning of the base product. This is one of the great advantages of the "plug-in" architecture - a new version of a stage type can be introduced without needing to create a new version of DataStage. And then, since Oracle preserves a certain level of backwar...

Being new, I suspect that the problem for the OP was simply one of not having the appropriate terminology ready to mind. I believe the question (uncertainty) has been resolved.

Try dragging jobs from the Jobs branch of the repository directly into the job sequence. These seem to be OK. There also seems to be a problem where the job's category is more than one level deep. I don't know if this is related. And, as noted above, I did not fix this systematically, so I can't gue...