Mailing-List: contact java-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: java-user@lucene.apache.org
Received-SPF: neutral (asf.osuosl.org: local policy)
Message-ID: <49658.222.152.67.193.1131669008.squirrel@msqr.us>
In-Reply-To: <Pine.LNX.4.58.0511092304340.10218@hal.rescomp.berkeley.edu>
References: <49742.222.152.67.193.1131580094.squirrel@msqr.us>
    <Pine.LNX.4.58.0511092304340.10218@hal.rescomp.berkeley.edu>
Date: Thu, 10 Nov 2005 16:30:08 -0800 (PST)
Subject: Re: efficiently finding all terms used on a particular field
     withinDocuments matching a query
From: "Matt Magoffin" <apache.org@msqr.us>
To: java-user@lucene.apache.org
User-Agent: SquirrelMail/1.4.3a-12.EL4
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Importance: Normal

> : I'm wondering if there a more efficient way to accomplish this?
>
> I believe there is -- provided the terms are index.
>
> 1) Get yourself a BitSet representing the Documents you are interested in
> (you mentioned having a a date range, you can either use a RangeFilter nad
> call the bits method directly, or you can do a search using a
> HitCollector)
>
> 2) Look at the code that acctually makes RangeFilter work.  It iterates
> over a TermEnum between a low and high value.  for each term it finds, it
> uses a TermDocs to record the docid.  You could do something very similar,
> looping over all terms in the field you want.  but instead of recording
> the docid, add the term to your Set object -- if and only if one of the
> docids from the TermDocs is in your BitSet from step #1 above.
>
>
> ...that should be faster then the appraoch you have now.

Thank you, I had thought a BitSet was appropriate here somehow, I'll work
on this approach.

-- m@


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