1- Searching like something* is the easiest way. 2- I don't know if you have some danish stemmer that can split the Compound words or not, but the best way would be to write an analyzer that can do this job.
xxxyyyzzz ----> analyzer ----> xxx yyy zzz. If you use that analyzer both in indexing and searching then you don't have to do nothing. Ex. When indexing xxxyyyzzz , tokens will be xxx yyy and zzz. When searching xxxyyyzzzz , tokens will be again "xxx yyy zzz". Besides that searching xxx or yyy (or zzz) will also return results. DIGY. -----Original Message----- From: Thomas Skovsende [mailto:[EMAIL PROTECTED] Sent: Friday, October 19, 2007 12:47 PM To: [email protected] Subject: Problem with words that have been split Hi guys, you were a great help last time, and I have a little problem. In my current setup lets say I have a tekst i want to find: "SomethingElse" but my user happen to search for "something else". Now - what I do to modify their standard search abit is that i actually rewrite their search to: "(something OR something~0.6) (else OR else~0.6)" Problem is that this does not hit on SomethingElse. It is a particular nasty problem because alot of words in danish are put together to form one word, but alot of people split it up in several words due to influense from English. Best regards, Thomas
