[Xapian-discuss] Word combinations

Garrett Smith g at rre.tt
Thu Dec 31 01:48:26 GMT 2009


If I'm looking for "magicjack", I might enter these search terms:

 "magic jack"
 "magicjack"
 "jack magic"

The content I'm indexing might have either of the first two variants
(probably not the second).

Without knowing these terms and their various word combination
permutations up front, what's the recommended approach for indexing
and searching for this scenario?

My guess is that you'd use a dictionary to split combined words into
parts (e.g. "magicjack" -> "magic" + "jack") for both indexing and
searching. If this is the right idea, is there any help for this built
into Xapian?

Garrett



More information about the Xapian-discuss mailing list