[Koha-zebra] Searches failing because of bad data

[Ted Romer]

I'm a Koha newbie trying to set up a small deployment for our school.  I
managed to import some MARC records incorrectly, with the effect that some
searches fail with "utf8 "\xC3" does not map to Unicode does not map to
unicode".  I added some debug statements to identify the specific records.

I'd like to avoid starting over from scratch, so I tried deleting the errant
records from the catalog (by doing a batch item delete via Koha, and
deleting the rows from the biblio table for good measure), and dropped and
rebuilt the Zebra biblio index following
http://koha-community.org/documentation/faq/searching/.

Any thoughts on where the data that I'd like to get rid of is lurking?

Thanks,
Ted
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </pipermail/koha-zebra/attachments/20110913/0d6da16c/attachment.htm>