DSXchange

Without seeing your job design, I don't know whether this is appropriate advice. However...

By default, a sequential file stage maps "" to NULL. This might be the source of your nulls. You can alter this default behaviour on the Format tab.

If you read my response to your original post on this topic, you will note that I told you what you were doing wrong, and predicted that this error message would appear.

At UNIX, does the user executing the DataStage job have execute ("x") permission to sendmail_testseq1.sh ?

If not, use chmod a+x sendmail_testseq1.sh to ensure that this is the case.

Ray - I thought active stages like the Transformer stage were always separate processes. Is this only the case when row buffering? Not necessarily. When two Transformer stages are connected by the one link, they can be called as subroutines from the same process. This is why the OP's approach worke...

This suggests that there's an undocumented limit in the Importer on the number of table definitions it can queue up - or it may be a limit on the number of lines in the control that displays them. In either case, applying a filter means less work when retrieving the list of table names from the data...

Roy's got it. If you try to run this as an after-stage routine you may have more success, since the after-stage routine runs in the same process as the transformer stage itself. The STAGECOM.STAGE and STAGECOM.PIN arrays are declared to be in COMMON (in the header file DSD_STAGE.H) and COMMON variab...

Welcome aboard! :D Without knowing the structure of SMF records, the best advice I can offer is "probably", but you may need to pre-process them. The mainframe can FTP them to your DataStage server; hopefully you can obtain the metadata from a copybook or DCLGen. The ComplexFlatFile stage is probabl...

Take a look at the code generated by compiling a Transformer stage (in the RT_BPnnn directory on the server). You will see that the column names are not actually used; pseudo-names like Column%%3 are used instead. This allows column names imported from actual tables, but which don't conform to the r...

When your hashed file is approaching 2GB, the command is: RESIZE hashedfilename * * * 64BIT USING pathname to convert the hashed file to 64-bit addressing and, therefore, the ability to be larger than 2GB without breaking. The USING pathname refers to a directory where there's enough room to build a...

It's actually not the whole story; there are extra properties later in the record ("extended properties").

However, for most purposes, Kim's approach is sufficient.

I did create an I-type for decoding the SQL data type, will post this once I find it!

Create a job - even a server job - to perform the task, and run that job. Can you please explain why you think that you have to do this in a Before/After Subroutine? Isn't "beforehand" sufficient? That means you have the option of a batch, job control routine or job sequence to execute the one job (...

The best place to learn UniVerse is from IBM. Most countries outsource UniVerse training; contact IBM to find out to whom in your country. Kim (anyone else reading this beware of brain implosion!) WHEN clauses, if used, follow and are constructed like WHERE clauses, and precede GROUP BY clauses. Unl...

I missed it first time; the original post stated that row buffering is not used. Please check that this is so (a) project wide - set in the Administrator client and (b) in the job properties - set in the Performance tab.
Other than that it's hard to imagine what might be causing the issue.

Are any other errors or warnings reported/logged? What volume of data (approximately) are you trying to process? Do any other applications on that machine generate memory errors? Have you logged a call with your support provider?

Welcome aboard! :D 1. On a Windows platform DataStage can call ActiveX (OLE) functions that have been installed on the DataStage server. A DataStage interlude routine (automatically created) is used. In DataStage help, find the topic External ActiveX (OLE) Functions. Here is a partial quote: To use ...