That is wonderful, working great. Thank you. The data I am dealing with fortunately has no '<' characters in the literal text. And I have a couple of working parsers from CPAN implemented but for this simple task the regex is a better tool for the job I think.