Mailing-List: contact user-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@couchdb.apache.org
Received-SPF: pass (nike.apache.org: domain of nicholas.evans@gmail.com
 designates 209.85.216.47 as permitted sender)
MIME-Version: 1.0
Sender: nicholas.evans@gmail.com
In-Reply-To: 
 <CABvT1DG77zjNSWCRmYQbNhu6OfuSenKsGFUa+JPiesJzJe0P6A@mail.gmail.com>
References: 
 <CACV2Mxd0A7Jii5jc=sHdK5x1CkqANLm6AjaB_9o+R_-sa7GKEg@mail.gmail.com>
 <CABvT1DG77zjNSWCRmYQbNhu6OfuSenKsGFUa+JPiesJzJe0P6A@mail.gmail.com>
From: "nicholas a. evans" <nick@ekenosen.net>
Date: Tue, 27 Mar 2012 13:28:47 -0400
Message-ID: 
 <CACV2MxfXeNWvq_hxSFeR6CDR_m7UnEC+uEw7LNvORf2eLzidhg@mail.gmail.com>
Subject: Re: Sorting CouchDB data using couch-lucene
To: user@couchdb.apache.org
Content-Type: text/plain; charset=ISO-8859-1

On Tue, Mar 27, 2012 at 1:14 PM, Robert Newson <rnewson@apache.org> wrote:
> cross-posting is generally bad form as it splits the conversation. If
> I answer here or on SO, someone potentially loses out.

Good point.  Sorry about that.

> That said, couchdb-lucene is an independent index, sourced from the
> database. As such, it has no access to the reduced values nor does it
> have an equivalent feature. Where did you see that suggestion?

I saw that suggestion at http://stackoverflow.com/a/2821476/110681

I think the idea wasn't so much that couchdb-lucene would read from
the view as much that couchdb-lucene would give me a completely
different way to approach the problem.  Is there some way for lucene
to index the data so that it could give me ranked results?  E.g. if
every document has a doc.sender, is there some way to store the
senders in lucene and then ask for a ranking of senders with their
counts?  I don't know enough about lucene or full-text indexing in
general, so I hoped and assumed there might be some way to twist the
lucene index to do my bidding.  :)

If lucene can't help out and no one knows of an open source chained
map-reduce views implementation, then the polyglot persistence
approach sounds best to me, since redis zsets are perfect for this
particular task.

Thanks.
-- 
Nick