Whoops, meant to send this to the list, not just Jordi.
---------- Forwarded message ---------- From: Jon Gorman <jonathan.gor...@gmail.com> Date: Mon, Jul 13, 2009 at 9:07 AM Subject: Re: ERROR : "\xF9" does not map to Unicode To: Jordi Pallares <jpalla...@cbuc.es> > I try to convert to marcXML record with yaz-marcdump and I works perfectly. > I see the marcXMl record and I found an error in the field 008 after the > country code I found a strange character. I don't know how it appears > here but I found some in a few part of my catalog. It doesn't look like valid utf-8. Utf-8 doesn't have a single f9 character. (It will end being a two byte code, but that doesn't look like the case here. Most of the characters are I think you may need to go back and step through the conversion process and see what the very original source record has in that part of the header. (Look at it with a hex editor there around character 296 or so (128 hex). Something seems to be going wrong somewhere in the conversion process. If the original records are mucked up, there may not be anything you can do besides replace it with some other character. Given that it's in the leader I suspect you'll want to simply replace it with a space. How to do the latter the best depends on the frequency of the error, your records, and your system. Jon Gorman
