It gets better than that. The reason I need them in CSV format in two columns of origin,destination like that, is because I have to import them into... wait for it... a Microsoft Access .mdb file, so an ASP page can query it. The import only happens rarely, but often enough to want to optimize the tool that is doing the conversion.

I'm slowly converting the system (carbon-based and silicon-based) to perl/mysql/php, depending on the needs, but for now, it has to remain in ADO/.mdb format during the transition.

I'll take a peek at Tree::M though and see what that can do for me. Thanks.