Sanitizing data

I've got an interesting problem that I'm hoping to solve with XSLT.

Basically I have an XML document inbound which could contain additional nodes that I do not currently handle in my XSL template; or worst still a node with multiple child nodes of which I am only expecting a single child node.

Can anyone give any pointers on the easiest way to trap and identify the additional data in the source XML that my template does not handle?

I've considered:
* Using XSL debugging commands to throw exceptions, but how to identify them?
* Pumping out all un-processed data to a single text block at the end of the FO for visibility. But how can identify un-processed nodes?
* Pre-parse the xml with another XSLT, kind of redundant/not ideal

Thanks, JV