DSXchange

I would never force this type of error - doing an operation on a nullable field. Add IsNull() or similar handling to your job; link1 should have the constraint "ColA=1 AND NOT(IsNull(ColB))", link2 should be "ColA=1 OR ColA=3". To your issue, what happens when the output of link2...

jgreve - thanks for the post. I, too, have had my issues with vectors and decided not to try using them. I still need to go back to the column export stage and see if a DS schema could somehow be used to generate the columns that I want, but I think I'll opt for the transform stage route as you've d...

Since you are in a sequencer job things are a lot easier than in a PX job directly. The KeyMgtGetNextValue routine can be called directly from a routine activity to give you a unique number, then pass that to the job that does the actual processing as a parameter to get your unique file names.

The difference betweena manual start in the director/designer and a scheduled one is usually in the userid which runs the job, and an external C object would cause that type of error if the read-execute permissions are not set correctly.

I have a VarChar() string coming from a MQ-Series input stream. This field contains a COBOL "OCCURS 0 TO 20 TIMES DEPENDING ON" structure of PIC X(10) with the first two positions being the depending on number of array elements, for example "02HELLO GOODBYE " or "04GOODBYE C...

In DataStage, what data type is this column? I ask because date data types have no format, but string types do. A string would then need to be converted to the data type that Oracle is expecting.

Are you calling a user-written function in that transform stage?

The space available for use in a dataset is limited by the space in the directories specified by the APT_CONFIG file for that run. When that limit is reached the DataStage aborts with a clear error message, something like "disk full". The "orchadmin" command allows you to list th...

Good news there. I assumed that PEEK would work with multibyte data, and now have learned not to trust that stage completely, either. I prefer text files anyway, as the contents are easy to manipulate and view in hex or other modes.

Take things a step at a time. Make a job with just your MQ stage and a text output file or a PEEK stage. Does the XML (as a string) get correctly processed through the MQ stage and output? Once you have that working, add the XML input stage; if it no longer works you know you don't have a MQ problem...

Even with RCP enabled, you need to explicitly declare your modify stage modified output column and need to enable the "microseconds" in the extended attribute.

I mean go into the designer, open up the job in question and do a "force compile" or normal compile on it. Does that work?

Compilers on some platforms cost $$$ so it makes sense not to put a compiler in, plus it does ensure that no changes can be made in production.

nvuradi wrote:I have given all the required information for the questions...

I see that you went back and re-edited an earlier post to add that missing information.

Do you have the microseconds enabled in your output column under the "extended" tag?