DSXchange

What you really want is a slightly different job design. "Insert New Rows or Update Existing Rows" is not efficient for updates; an insert has to be attempted and fail because the row already exists before the update is attempted. A better design is to have two links into the database, one doing "in...

Expressions in DataStage are exactly that: expressions. Expressions generate values.

They are not assignment statements, and therefore should not contain assignment operators.

Volume is irrelevant.

You control what the DataStage job does by designing it.
Direct the rejected rows to a text file for maximum "efficiency" if this is a concern.

The user dsadm can be a DataStage administrator, but you must install as that user. In version 7 you can also configure the server with "impersonation mode"; read the caveats in the uvconfig file.

You also have to specify what you want to happen if there isn't a four digit number in the input column. Also, what do you want to do if the CHAR(5) column contains five digits? You can use pattern matching to determine whether there is a four digit number. For example, you might set up the followin...

It's not PIC BASIC and it's not PICK BASIC. It's DataStage BASIC (or DS BASIC if you must persist with these infuriating contractions).

In the header of the DSX file is a Character Set setting.
You can try changing this to the appropriate value for your Italian environment then try importing.

You need more money.

It's a separate component for DataStage, not part of server edition. These days Ascential call it DataStage EE (Enterprise Edition). But we here seem to stick with Parallel Extender, especially since the design canvas still carries the word "Parallel".

I think the time has come for you to do some diagnosis. Try connecting to the data source with the DS_CONNECT command in your DataStage project. Within the branded_odbc directory is a READ.ME file that is worth reading through, and a demoodbc executable in branded_odbc/demo (and a readme file that e...

No. The sqlldr utility only does bulk loads. This is typically the final step in a DataStage job, which also performs extraction from source, cleansing, transformation, aggregation, and so on. You're not comparing apples with apples here. However, parallel extender (PX) does give you potentially hug...

That looks OK. I presume SQL Server is listening on port 1433?

One of them. This is the only safe answer. It depends on how the combination of ODBC driver and database server resolve your query. There's nothing to stop you specifying which one you want by switching to user-defined SQL and adding an extra constraint, even a correlated sub-query. Of course your l...

"Synthetic" means generated by the underlying "UV" engine (now DataStage Engine). It's like a system variable; you don't have to do anything; it's just there. To create a column in a UV table as multi-valued, specify the keyword MULTIVALUED in the column definition. For example: CREATE TABLE MyTable...

No. The sqlldr utility only does bulk loads. This is typically the final step in a DataStage job, which also performs extraction from source, cleansing, transformation, aggregation, and so on. You're not comparing apples with apples here. However, parallel extender (PX) does give you potentially hug...

This one is tricky. Default precision is 15 but should be able to be modified with the EXACTNUMERIC configuration parameter. Execute the command smat -t | grep EXACTNUMERIC to determine whether your change was effective. Try editing the JOBCONTROL.H file (in the DSINCLUDE subdirectory), adding the f...