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 alon.keren@gmail.com designates
 209.85.212.52 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <20120415061323.GD8135@translab.its.uci.edu>
References: 
 <CABzyt1G-19VR-To-ph=RqjYTUN0xZEFGs3TzJU8bO_ExsJeqtw@mail.gmail.com>
	<20120415061323.GD8135@translab.its.uci.edu>
Date: Sun, 15 Apr 2012 12:00:38 +0300
Message-ID: 
 <CABzyt1HB_VD8n31qCQuBdcS-h2DGANMqwx+KRO+dfqHJ8TUDSw@mail.gmail.com>
Subject: Re: Reduce just N rows?
From: Alon Keren <alon.keren@gmail.com>
To: user@couchdb.apache.org
Content-Type: multipart/alternative; boundary=20cf3071d0a27d787904bdb3f0e2

--20cf3071d0a27d787904bdb3f0e2
Content-Type: text/plain; charset=ISO-8859-1

On 15 April 2012 09:13, James Marca <jmarca@translab.its.uci.edu> wrote:

> CouchDB will compute reduced values for what you select.  If you just
> ask for values from A to B, it *only* will compute the reduced values
> over that range.  So you can get "clever" with the key value, using
> something like
>
> map: emit( [user,game,trynumber], score);
>
> where trynumber is some value that is guaranteed to increase with each
> completed game score stored.
>
> Your reduce could use the built-in Erlang  _sum
>
> Then you can just request something like...hmm
>
> startKey=[user,game,BIGNUMBER]&order=descending&limit=10&reduce=false
> (where BIGNUMBER is something bigger than the highest try number of game).
>
> This will give 10 values, and you can do the average lickety-split
> client side, OR you can do one query to get highest try number, then
> another to get between that game and ten back to let couch compute the
> sum for you.
>

Thanks!

I think a simpler alternative to 'trynumber' is the game's timestamp, and
BIGNUMBER could be replaced by '{}' (see:
http://wiki.apache.org/couchdb/View_collation). That's what I'm doing at
the moment :)
Unfortunately, as numbers of games and game-types grow, this would become
pretty demanding in CPU time and number of calls to couch.


>
> I'd play around with that kind of an approach, but I would recommend
> *against* computing the mean in the reduce, although you could do it.
> It is much slower than using _count or _sum.  Using my approach, Couch
> will get the sum value, not the average, but you also should know the
> rows and so can compute the simple arithmetic mean much faster than
> serializing the data back and forth from JSON on the server.
>
>
You're right, of course, and indeed my actual reduce is a sum operation -
the average is calculated from its final value.


>
> As an aside, invariably I've found that when I use arrays for keys I
> generally run into at least one case where I want to shuffle the order
> of the keys and therefore generate an almost identical copy of the
> view's data just to get a different organization of the tree.
>
> Hope that helps,
> James
>

>
> On Sun, Apr 15, 2012 at 02:02:21AM +0300, Alon Keren wrote:
> > Hi,
> >
> > I have the following scenario:
> > a player can play one of many types of games. After finishing a game, the
> > player's score is kept in a new document.
> > I'd like to list a player's average score in the latest 10 games, for all
> > types of games.
> >
> > I don't see an easy way to solve this with CouchDB, unless there was a
> > feature that limited (group-)reduce operations to just the first 10 of
> all
> > possible rows.
> >
> > Has anyone else encountered a similar situation? Would this feature be
> > useful to others? If so, would it be technically feasible to implement?
> >
> > Thanks,
> >
> >   Alon
>

--20cf3071d0a27d787904bdb3f0e2--