DSXchange

What has changed between "previously" and "now"? "Nothing" clearly is not the correct answer.

Not 100% sure whether that's true or not, but if it is you can put the Sequential File stage into a server Shared Container to make use of this particular functionality in a parallel job.

You might also investigate an approach using the Schema File property of the parallel Sequential File stage.

Library search path is an environment variable.

The library path of a routine ought not to change (probably). Does the routine definition complain if you omit it, and rely on the shared library search path to find the routine at run time?

You don't have to script anything - use QualityStage designer to design your input pattern override. You don't have to investigate - this step would, however, allow you to see what other patterns or text strings are in need of work. Typically you would, however, investigate the results of a standard...

Someone tightened up file system permissions on the server?

Please explain what these environment variables do, and how you came to decide on these specific settings for them, so that we all might understand your resolution and why it may have worked.

Stick them in a spreadsheet and add them there. What results do you get?

Welcome aboard.

There is none. Parallel jobs explicitly forbid a passive stage having an input link and an output link, because this obstructs pipeline parallelism.

In a parallel execution environment you must implement this design using two jobs; one to write and one to read.

Define 105 columns. Create "missing columns" rules in the table definition for the Sequential File stage.

You do not connect to a DataStage server in version 8. You connect to Information Server (via port #9080 by default) where you are authenticated. If you are authenticated, the login dialog for Designer or Director offers a drop-down list of server/project pairs to which you can connect. Similarly Ad...

That's why you got 216 rows out.

If you must create a Specification, it looks like

The issue is one of timing. Your approach will only work if you use auto-commit (rows/transaction = 1). Otherwise the insert may have been sent but not yet committed when the update is sent, so the update is rejected because the row has not yet been inserted.

Welcome aboard. I don't know that there's a conventional format string for this timestamp format. I'd try reading it as VarChar and modifying it to a conventional timestamp format (adding 12 to the hours if the designator is PM, changing the decimal placeholder from ":" to ".", changing MDY order to...