Folks,
I would think this would be basic/simple but I cannot get it to work. I am trying to logically break up my workflow into a series of subjobs that can be reused easily. My overall workflow is to retrieve 2 files (each one from a different MQ) then parse each file as fixed width delimited and then finally join the two files together into one result dataset based on a key common to the content of both files.
I have broken out the parsing of a file to a child job - in there it contains tFileInputPositional - if I use a tLogRow component within the subjob I correctly see each row being output to the logs. However, if I pass the tFileInputPositional directly to a tBufferOutput for return to the parent job, nothing ever gets returned - I.e. back in the parent if I put a tBufferOutput from the output of the subjob, I get no rows back.
Is tBufferOutput designed to handle multple rows? If so, can someone give me an example of this at work. If I am using the wrong component, what component should I use?
Yes, tBufferOutput is used to returns data back to parent job, you need to define the same schema as tBufferOutput's schema on tRunJob which is calling the child job. eg:
child job
tFileInputPositional--tBufferOutput
parent job;
tRunJob--main--tLogRow
In this example, you will see the result are printed on the console.
Best regards
Shong
Shong,
Thanks for replying and that IS how I have it set up. However, I've solved my pesky problem: Under 'advanced settings' for the tRunJob that represents the child job in the parent, I had to ensure 'propagate the child result to the output schema' for the results to return. You'd think that *duh* I would have known this but the thing is, there are several times I have used sub-jobs (such as processing data off MQ, etc.) and I have NEVER had to check this before to get my data back to the parent. Which makes this a confusing inconsistency for me if anyone is able to explain?
thanks,
D.
