Thank you for your reply. That is roughly the approach so far. Identify the source and target column - make some sort of notation - very similar to your @swap array - and let it drive OLE operations. It is simple but somehow I feel I am not doing it right. Every time I need something more sophisticated I need another new class of "rule" and associated handler code. On a positive side that rule syntax is primitive enough that anybody can maintain the rules file.

I have not done a lot in XML but I thought that perhaps the way to go was to convert both XLS files to XML and then use something like XLST to manipulate target XML based on source one. Is it something that offers some inherent advantage or am I opening Pandora's box trying it?