kael@email.arizona.edu wrote:

...

Han, Yan wrote:

...

That is right.

Look at this

http://arrakis.library.arizona.edu:59/cgi-bin/koha/opac-detail.pl?bib=10 48

...

You can only search "22269; | 20013;& | &"

Kia Ora ano

Further to my last email, do you have access to the Koha mysql database. If so can you please do a select * from biblio where biblionumber=1048;

Is that where the 22269; is stored? or is it in the marc_word table?

When I do the SQL query on the on the biblio table I get (tranfered to a hopefully email-friendly layout): +---------------+------------------+ | biblionumber | 1048 | | author | 韩 | | title | 中国 | | unititle | | | notes | NULL | | serial | NULL | | seriestitle | | | copyrightdate | NULL | | timestamp | 20051006163507 | | abstract | NULL | +---------------+------------------+

The '22269;' is also stored in the marc_word table: +----------------+--------+ | bibid | 1048 | | tagsubfield | 245a | | tagorder | 11 | | subfieldorder | 1 | | word | 22269; | | sndx_word | 2000 | +----------------+--------+

Ah ha. So the biblio table is storing the entity properly, eg title 中国 So when the record is displayed it displays just fine. But marc_word (which is what is searched on) has had the &# stripped. So it looks like its a propblem with the routines storing data in the marc_word table. Ill look into the routines that do the storage and look for a fix. Anyone else who knows more about the MARC tables and/or unicode want to yell out now and tell me im on the wrong track? Or where to start looking? Chris -- Chris Cormack Katipo Communications Programmer www.katipo.co.nz 027 4500 789