DSXchange

This is an entirely new problem. We gave solutions for the question you originally asked.

The new problem is solved completely differently, most probably by having the Sequential File stage generate a file name column.

Any way you like.

You can add leading zeroes to the one, or remove leading zeroes from the other.

Techniques for both can be found by searching DSXchange.

It won't affect your join result, which should be the same as if you did it using SQL in the database.

The same goes for data types - they must be compatible. Identical is best.

Welcome aboard. You must have an ODBC connection to SQL Server - since SQL Server does not run on UNIX. Then you can import using the ODBC Connector Import Wizard, or via the Orchestrate orchdbutil mechanism. But the first thing you need to do is to configure the ODBC data source name on the UNIX ma...

If you are using a Join stage it is mandatory that the inputs be sorted on the join key(s). If you use ORDER BY in Oracle use a Sort stage to specify "don't sort (previously sorted)". Partitioning is a completely separate question. Both (all) inputs to the Join stage need to be partitioned...

Have you installed Fix Pack 1? There are lots of things fixed in this. (You can read what they are at Information Center, you can download FP1 from Fix Central.)

My personal philosophy is never to try to diagnose messages from APT_CombinedOperatorController.

Disable operator combination and try again. That way you will ascertain precisely which operator threw the message.

The actual field widths must be recorded in the Display Width field of the table definition. Use the Sequential File table definition importer and specify Fixed Width as the format and enter the field widths right there. Then use the imported table definition in your job.

"Terminated unexpectedly" can mean that someone or something has killed a process or a network connection. Is the problem reproducible?

Are there any warnings in the job log? Is there a "Finished" event in the job log?

There's not 100% overlap, not least because Oracle "standards" are different from ANSI standard. UniVerse is ANSI-compliant as far as it goes, but has extensions unique to itself such as multi-valued columns.

Manually, using the Data Set Management utility.

There's nothing even semi-automatic available in version 7.

This information is deliberately not documented by the vendor.

Your best chance is to search DSXchange.

... or even in the ls command itself. Let's imagine you're using an External Source stage. The command is something like

Because you have bad data - data that does not conform to your specified metadata. Either get the data changed so that they conform to the metadata or change the metadata so that they conform to the data.