On May 1, 2008, at 4:09 PM, Jay A. Kreibich wrote: > As the document referenced above explains, you can create a user > function called "like(A,B)" that will over-ride the built-in behavior > for LIKE. If you're only dealing with one language, such as > Brazilian-Portuguese, then you can just custom code the various > accented characters used in that specific language.
Alternatively, one could normalize (i.e. transliterate) both data and query before hand. E.g. using Sean M. Burke's Unidecode [1] or such [2]: ação -> transliterate -> acao AÇÃO -> transliterate -> acao Москва́ -> transliterate -> moskva Ἀθηνᾶ -> transliterate -> athena 서울 -> transliterate -> seoul 北京 -> transliterate ->beijing For example, searching for 'nino' or any variation thereof, would match 'Niño': http://svr225.stepx.com:3388/search?q=nino http://svr225.stepx.com:3388/el-nino-southern-oscillation Cheers, -- PA. http://alt.textdrive.com/nanoki/ [1] http://interglacial.com/~sburke/tpj/as_html/tpj22.html [2] http://www.gnu.org/software/libiconv/ _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
