Mailing-List: contact java-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: java-user@lucene.apache.org
Received-SPF: pass (hermes.apache.org: local policy)
From: Paul Elschot <paul.elschot@xs4all.nl>
To: java-user@lucene.apache.org
Subject: Re: mutiple index question
Date: Fri, 20 May 2005 20:59:57 +0200
User-Agent: KMail/1.5.4
References: <A955EA1F8FE31749AEC8C998082F6C7C41A79F@hai01.hippo.local>
 <200505201411.10554.paul.elschot@xs4all.nl> <d6krnq$96t$1@sea.gmane.org>
In-Reply-To: <d6krnq$96t$1@sea.gmane.org>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200505202059.57179.paul.elschot@xs4all.nl>

On Friday 20 May 2005 16:21, Robert Newson wrote:
> Paul Elschot wrote:
> > On Friday 20 May 2005 13:58, Max Pfingsthorn wrote:
> > 
> >>Hi!
> >>
> >>I was wondering if Lucene has any sort of functionality to distribute 
> > 
> > indices so that different fields are stored in separate indices but they 
> > still refer to the same document. This would be great for a situation 
where 
> > there are many large documents which have frequently changing properties. 
> > Content and properties could be indexed separately. Even different sets of 
> > properties could be combined in maybe different MultiSearcher instances to 
> > speed up querying...
> > 
> >>Any ideas on this?
> > 
> > 
> > Lucene doesn't have this, so you'll have to split your docs over two 
different
> > Lucene indexes and adapt the search accordingly. Cached filtering helps a
> > lot, but setting up a filter can still be costly.
> > 
> > Regards,
> > Paul Elschot
> 
> Isn't this achieved by the new ParallelReader in CVS?

Without having had a look at that code: probably not, because
a reader does not know which field is the primary key.

The problem with the two indexes above is that a single document has
different fields in both of them.  Usually a search would start in the
"properties" index, retrieve the found primary keys from there, 
then use these primary keys to set up a filter in the "content" index,
and then repeatedly search in the content text limiting to this filter.

Regards,
Paul Elschot


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org