Best method of finding the largest numeric value in an xml document

I have an BLS routine where I need to find the largest numeric value in my XML document. It could be in any row or column. I'm not sure what column it will be in, so the aggregate function is not a good choice (unless I recursively call it for each column and then dynamically assign it to the output) I'm thinking of a couple techniques but I wanted to get some other's opinions first. In the past for things like this I've just nested a couple repeaters but maybe there is a more efficient way.

This is a standard Illumdoc XML format.