DSXchange

Is this really a server job? Are you using row buffering?

You must have seven columns in your job design metadata for the INSERT link, because your statement includes seven parameter markers (:1 through :7). And so on.

DB2 9.5 can also produce XML, which may be an easier solution. And you could use it to extract the data from the version 8.2 database.

Your PAT file lacks a STRIPLIST in the PRAGMA section, lacks a POST actions section (which is OK, you may not want any), and lacks an EXIT statement to prevent falling through into the subroutine. You have not shown examples of the kinds of data you will be processing, but any with AARON, ABBEY or J...

Moderator: please move to QualityStage forum

Have you had any QualityStage training? The detailed description you request is about 50% of the QualityStage users' guide. There's no way we're going to re-hash that here.

That's how I would do it too.

OK, then try something like

The output data type needs to be string of some kind. You can not prepend zeroes to a numeric data type - they are stored internally in binary format.

What are you using to view the data?

As I already answered in one of your other posts, use the Compare() function if you need to force a string comparison. Syntax of the Compare() function can be had from the DataStage BASIC manual or from the on-line help.

HOW (with what function) are you converting the string to an integer? Are you totally confident that each input value CAN be converted to integer? Do you check?

Modify stage can create new column but the value must be based on an input column using one of the functions available to this stage type.

What about dssearch, dssh, dsexport, to name just three?

Use the log processing functions (DSGetNewestLogId(), DSGetLogSummary(), DSGetLogEntry()) in the routine or the equivalent options in the dsjob command.

So what? It still works. "Fiche" = file, "Ligne" = line.

The words of the script language (like "for", "set" and so on) are universal.