DSXchange

The easiest thing to do, especially if you are already consuming the entire string and passing it as a column value, is to just "zap" it. Pass the string thru a Basic Transformer and either pull off each of the pieces or just use something like ereplace and blow away the whole schemaLocati...

Hard to say...certainly the fact that the Job runs on the other machine means that everything is coded correctly (assuming you just moved the .dsx).....where are you getting your input from? any chance it is getting bogus data in this case? What other differences are there in the environments beside...

Look in the documentation, and then the Stage or Connector for all of the various "Data Element" options that you have. There are a whole set of them for gathering bits and pieces of the MQ header. I can't remember exactly what it is and am not looking at DataStage right now, but assuming ...

If the xml content is actually "in" a column coming over on the link from the dataset stage, then be sure, in the xmlInput Stage, to check the box that says "xml content". It is not the default. .....sounds like it is taking whatever string is in there and trying to use it to loc...

One giant element? The whole document? Send the chunk of file (read it with the folder stage in a Server Job for the simplest scenario) as one column into a Transformer and just concatenate the few xml header and tag information to the front and back of the column. No sense in spending time on an xm...

I can only guess at the moment....it certainly seems like you have included the necessary methods. Do you have any simpler samples that work all the time? There is sample code at my blog (see below)....does that work? When you say it works "sometimes".....what is different? One "maybe...

I am not able to look at it right now, but I suspect that the filter style list feature (the little funnel icon at the upper right) may work against a general term list in the dev glossary and allow you to save things in .csv. Perhaps not --- I'll have to try it. Not a great solution if it works, bu...

At this current time the Export facility only looks at the "Published" Glossary.

Ernie

ya. I know...your're right...it's there. ....I wasn't clear. ...I meant "hard to tell without seeing it in the gui"....the xml Stage is weird like that --- the visual appearance of the xsd after being imported to the library (with it's various icons for each node type, etc.) is very helpfu...

I "think" what you are trying to do is simply pass the value thru the Stage....whether it was hard coded or pulled from the External Source, there is a string value on the input link to the xml stage. In order to "pass it thru", you have to formally map it. [the new xml stage doe...

Without seeing the xsd it is hard to say, but if the address node is a maxoccurs=1, then you should be able to use an Aggregate step for that node up above and then that will satisfy the Output step mapper.

Ernie

As for your second question, use a parallel Job if you can. In Server, the xml Stage only can have one input link. In a parallel Job it can have as many input links as you need.

Ernie

In general, you should be able to get the address and the person info in the same mapping step, but it might be the design of your xsd that is stopping you..... Check the address node and see if it is allowed to repeat (maxoccurs unbounded or > 1)....if it is, then these are entirely "independe...

by the way...this is in the wrong forum...let us know if this is a Server Job or a Parallel Job so that it can be correctly moved.

Ernie

As you noted, you will probably need to use the xmlInput Stage "after" the Web Service Stage.....but first just make sure you are calling the service correctly. Copy your Job to a new one, and delete any xml stages that follow the web service stage. Have only a single link coming from the ...