Mailing-List: contact couchdb-user-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: couchdb-user@incubator.apache.org
Received-SPF: pass (athena.apache.org: domain of paul.joseph.davis@gmail.com
 designates 64.233.178.246 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=message-id:date:from:to:subject:in-reply-to:mime-version
         :content-type:content-transfer-encoding:content-disposition
         :references;
        b=obQHfKHo2UIfOFdqJ81W6GfOCdBSHauihRfeHUu4XpMW+etkoKo+cIvpl69deSaETD
         J11RtOWNR2iZ3LlusmEIy7ncxTb0BIUK8BiLPbUGUJs95m9FJEZVqGoqOn33xhqGIizr
         VQRyW1Fo5tEWG9TH2lGmpJnDi8hydmP+LCCc8=
Message-ID: <e2111bbb0807090902j45f7d280g1d52288341080dab@mail.gmail.com>
Date: Wed, 9 Jul 2008 12:02:50 -0400
From: "Paul Davis" <paul.joseph.davis@gmail.com>
To: couchdb-user@incubator.apache.org
Subject: Re: thoughts on bulk_docs
In-Reply-To: <372EC480-C40B-48EF-B6B1-1ABD436BF584@blit.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <372EC480-C40B-48EF-B6B1-1ABD436BF584@blit.com>

On Wed, Jul 9, 2008 at 4:10 AM, Troy Kruthoff <tkruthoff@blit.com> wrote:
> Regarding previous discussions about multi-key gets, I was thinking why not
> just bring bulk_docs into parity with the rest buzzword we're a giddy about.
>   Example, instead of
>
> {
>  "docs": [
>    {"_id": "0", "integer": 0, "string": "0"},
>    {"_id": "1", "integer": 1, "string": "1"},
>    {"_id": "2", "integer": 2, "string": "2"}
>  ]
> }
>
> do
>
> {
>  "put[||post]": [
>    {"_id": "0", "integer": 0, "string": "0"},
>    {"_id": "1", "integer": 1, "string": "1"},
>    {"_id": "2", "integer": 2, "string": "2"}
>  ]
> }
>
> so later, when supported we could do cool stuff like:
>
> {
>  "get": ["id1","id2","id3"]
> }
>
> and/or even cooler
>
> {
>  "get": [
>    {startkey:"a", endkey:"b", count:100},
>    {startkey:"c", endkey:"d", count:100},
>    {startkey:"e", endkey:"f", count:100}
>  ]
> }
>
> and/or even bulky crazy
>
> {
>  "get": [
>    {startkey:"a", endkey:"b", count:100},
>    {startkey:"c", endkey:"d", count:100},
>    {startkey:"e", endkey:"f", count:100}
>  ],
>  "put[||post]": [
>    {"_id": "0", "integer": 0, "string": "0"},
>    {"_id": "1", "integer": 1, "string": "1"},
>    {"_id": "2", "integer": 2, "string": "2"}
>  ],
>  "delete": ["id1","id2","id3"]
> }
>
>
> For me, keeping the concepts of rest exposed through the bulk_docs api just
> seems sexy.  I'd take a crack at this, but it would take awhile, as Erlang
> and I have yet to dance, although my beer goggles are starting to work, so
> it may not be long! -- Meaning Erlang sure is ugly, but couchdb (and
> ejabberd) are starting to make it look attractive (for me anyways...)
>
> -- troy
>

I'm suprisingly liking this idea. IIRC, there's already code in the
couch_view.erl to do this insert/delete/fetch in one pass through the
view which makes me think it wouldn't actually be that hard to
implement.

Any core developers want to chime in with a slap down or blessing for
attempting to implement it?

Paul