Hi,

It would be possible if you index tokens not as "words", but as "character 
ngrams".  You'd need a custom analyzer for that.  Code for character-based 
ngrams already exists in Lucene contrib, but you'd need to add it to Nutch.


Otis
--
Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch



----- Original Message ----
> From: Ben Chapman <[email protected]>
> To: [email protected]
> Sent: Monday, December 15, 2008 9:58:24 AM
> Subject: Re: Partial word searches?
> 
> Does anyone have thoughts/suggestions for partial word searches? Is it even
> possible?
> 
> -Ben
> 
> On Fri, Dec 12, 2008 at 4:26 PM, Ben Chapman wrote:
> 
> > Hi,
> > Is there a Nutch configuration option to enable search results for partial
> > word searches?
> >
> > For example, searching on Dec will return 0 pages, while searching on
> > December will return multiple results. I'm hoping that the search results
> > for Dec will include Dec* (December, Decedent, Decca, etc).
> >
> > I've read through all the documentation, and googled to no avail.
> >
> > Thanks guys.
> >

Reply via email to