Pipelining XSLT transformations

Hi,

I’m experiencing a strange phenomenon when pipelining multiple DXL transformations. I have an agents which creates a new document aggregating multiple existing docs from a different db.

The aggregation is the result of DXL exporting the docs while applying a first xslt filter, applying a second transformation to aggregate the doc and finally importing it. Now:

  • if I just pipeline the different steps I do not get some of the fields in the final document and some contain strange values;

  • if I save on file the exported and filtered xml stream, apply to it the aggregating transformation manually (e.g. using Saxon) and import the result, the doc looks just fine!

Questions:

  • does anybody have a hint on where the problem might be?

  • are there any known limitations to the pipelining mechanism?

Many thanks, Enrico

Subject: Pipelining XSLT transformations

Okay, I found the bug… in my code. The documents exported manually and those aggregated automatically where not from the same view, i.e., did not contain the same values. In addition I have now reduced the filtering and publishing steps to one single tranformation and everything works fine.

I really like the XML toolkit!

– Enrico