DSXchange

Even so, I suspect that Intel is preferred; there are references to Intel in the installation guide, and there is a specific reference to I386 in the folder structure.

TENACITY instructs Teradata to queue for retry if a utility slot is not available. Without out, the process simply fails if there is no utility slot available. By default, there are 15 utility slots.

Who is "u"? There's no-one of that name registered on DSXchange.

Any data source via ODBC.
Any text file.

You can also invoke QualityStage jobs from DataStage, in which case the sources in DataStage are all available; the data are streamed into QualityStage (without touching down to disk).

The rule actually says that the exception is "in expressions".

Anywhere you need a parameter reference in a property, such as a File property, then the surrounding "#" characters are required.

DSguru2B wrote:Only if you were sitting on a unix box. This would have been much easier.

How exactly?

Everything on DataStage on UNIX is available on DataStage on Windows.

Who is "u" ?

There's no-one of that name enrolled on DSXchange.

You query requested job name, category, object name and OLE type. The result contained job name, category, object name and OLE type. You have to trust that the database has selected only those records that contain "PROJDEF", since you have not asked for that to be displayed. QED If you want a differ...

Jump into the folder containing the Documentation Tool. Here you will find SQL scripts for creating a database. Open the doc_tool.mdb file using Microsoft Access (hold down the Shift key to bypass the startup macro, or press F11). From the menu, have Access generate a Relationships diagram. While n...

Primarily for the memory management aspects - there are no functions in DataStage BASIC for this. The STREAMS I/O module is used to increase performance of the Sequential File stage. Anywhere you see a subroutine call (for example when stage tracing) where the subroutine name begins with "$" a call ...

Oh, it will work (run without generating runtime errors) OK. It just won't yield correct results. But surely that's a problem for QA, not for developer! The solution using Convert is more efficient (processes the string fewer times) than the multiple Index() answer I gave earlier. Kudos to DSGuru2B.

From the command line you use dsjob for all job types. DataStage figures out which executable to use.

"Or" within an argument, as you have done, does not do what you expect. ("/"or"#"or"@"or"%") will always return 1, so that your Index() function is actually reporting whether any "1" exists in BISGROUP_ID. You need multiple Index() functions, with "Or" between those, one Index() function per charact...

Completely on the wrong track. ODBC is usually slower (a) because there are extra layers of software involved (includng but not limited to the ODBC driver manager and the ODBC driver), and (b) sometimes the ODBC way of doing things - particularly if the JET Engine is involved - is curiously ineffici...

The row counts reported in Designer are the rows being SENT to the database. Whether they can legally be loaded is another matter entirely. If the table does not exist then clearly the answer would be "no". And you should see lots of warnings in your job log in this case.