Various soundex algorithms implementations in 10g?

Hi,  

I've just had a query from a developer about soundex algorithms in Oracle 10gR2. I know about the soundex() function, which is an implementation of the Russell soundex (described by Knuth in TAOCP). The developer seems to think that there are some gaps in the Oracle implementation, though. (working on getting details/specifics of what he means by gaps in the implementation). Also, he's asked about some other algorithms, namely, Daitch-Mokotov and Double Metaphone. Anyone have any experience with either of these on Oracle? Anyone have PL/SQL implementations they'd be willing to share? J For now, I think I'm looking to code this in PL/SQL, and we won't be using any Oracle Text features.  

Finally, these are US Census names that we'll be searching, and from what I read, there's an algorithm called Jaro-Winkler which was specifically written for US Census name data, *and* Oracle has an implementation of that one in UTL_MATCH on 10gR2. So, I've asked the developer about this algorithm and whether he's considered it.  

I guess I'm just looking for any folks that have any experience implementing any of the above algorithms, any thoughts or things to look out for, pointers to public implementations of any of these implementations, etc.....  

-Mark  

--

Mark J. Bobak
Senior Database Administrator, System & Product Technologies ProQuest
789 E. Eisenhower, Parkway, P.O. Box 1346 Ann Arbor MI 48106-1346
+1.734.997.4059 or +1.800.521.0600 x 4059 mark.bobak_at_il.proquest.com <mailto:mark.bobak_at_il.proquest.com> www.proquest.com <http://www.proquest.com> www.csa.com <http://www.csa.com>

ProQuest...Start here.  

--

http://www.freelists.org/webpage/oracle-l Received on Mon Dec 10 2007 - 14:35:34 CST