Thanks. Your reply pointed me to an error in my thinking: I assumed that if a file contained accented characters and Japanese yen symbols that it had to be in UTF-8, not realizing that ISO-8859-1 could also support those.

What I'm really trying to do is copy tables from a production database into XML (using DBIx) and load it into a test DB using XML::Parser. XML::Parser, however, dies with an "invalid token" error message. Chasing this down is where the confusion arose.

Thanks again.