If I understand the PARSE command, you will set up a routine to roll through ALL of the tags. Each one will be passed into your routine in a special variable. The value of the data in the tag will be in another special variable. Based on the tag name, you can branch to a different handling routine to get the data value where you want it.
What's confusing me is that you say the XML is in a "file." Do you mean that the entire XML document is contained in a single record of a file? Or do you mean you have like 80-byte chunks per record?
Here is an example from IBM if you are dealing with a "chunked up" file: http://www.google.com/url?sa=t&rct=j&q= ... kQ&cad=rjt