Mailing-List: contact user-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@couchdb.apache.org
Received-SPF: pass (athena.apache.org: domain of b.candler@pobox.com
 designates 208.72.237.25 as permitted sender)
Date: Mon, 6 Apr 2009 16:00:55 +0100
From: Brian Candler <B.Candler@pobox.com>
To: Adam Wolff <awolff@gmail.com>
Cc: user@couchdb.apache.org, Tom McNulty <tom.mcnulty@cetiforge.com>
Subject: Re: Reduce Assumptions
Message-ID: <20090406150054.GA19604@uk.tiscali.com>
References: <e8d26ac40904031955j2b09debegcd351865c5d6f519@mail.gmail.com>
 <20090406085100.GA14371@uk.tiscali.com>
 <e8d26ac40904060746l3aaf7f13r83303b58a2f3c280@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <e8d26ac40904060746l3aaf7f13r83303b58a2f3c280@mail.gmail.com>
User-Agent: Mutt/1.5.17+20080114 (2008-01-14)

On Mon, Apr 06, 2009 at 07:46:37AM -0700, Adam Wolff wrote:
> I'm still not quite clear on
> the meaning of this:
> "reduce functions should not grow its output larger than log(n) where n is
> the number of input rows"
> 
> How is the size of the output measured? length of the JSON string? are
> the input rows
> the size of the value emitted by the map?

I don't know the details - I just copy-pasted that from the wiki.

> Does this mean it's wrong to
> have an entry in
> your reduce output for every map key?

I believe that's the case.

The reduce output forms a tree overlaid onto the document Btree: each Btree
node contains N documents plus a single reduce value from those N documents.
The Btree could be multiple levels deep, and you don't want to end up with a
single enormous value in the root node. But I don't know what that actual
size limit is.

Regards,

Brian.