Good point!

I'd give you the double plus if I wasn't 3 floors underneath the dungeon level!

I did ask if there was a 1-to-1 line concurrency, so you can infer from that that I was aware that the solution would not work if there was any mismatch in the line concurrency. Also adding .*? inside the left tag would catch any randomly added attributes, but yes, I can see why regex's are not the way to go if your less than 100% sure about the quality and consistency of the input data.

So that leaves me wondering just exactly how XML::libXML recognises the structure of a tag, what is the black magic therein? And also, how does libXML work on the file without loading it all into memory (the guy said they were extremely large.)