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 sebastiancohnen@googlemail.com
 designates 74.125.82.180 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=googlemail.com; s=gamma;
        h=content-type:mime-version:subject:from:in-reply-to:date
         :content-transfer-encoding:message-id:references:to:x-mailer;
        b=R10A+xfm8+sbM3ZNfYY9UeHMg8HnVcO++HH2NASoaVNI3Xk8L8j8XSKQUsdNfMfxwi
         sfDMjeHHudEN9BY7gOiegEnGyE4Ikf6xqPlwseYRYom2yNu3w0pxotvIY7IxOE6ryTzU
         U6p5cR6xLAbb4qX96mFQeisCUeMWmLaIeh6Kg=
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0 (Apple Message framework v1081)
Subject: Re: expected behaviour of changes feed
From: Sebastian Cohnen <sebastiancohnen@googlemail.com>
In-Reply-To: <4C5D7408.6060304@aol.com>
Date: Sat, 7 Aug 2010 18:52:10 +0200
Content-Transfer-Encoding: quoted-printable
Message-Id: <1BEF5DD1-EBA2-49F3-8B9A-67ECE5B7E7C5@googlemail.com>
References: <4C5D7408.6060304@aol.com>
To: user@couchdb.apache.org

hey cliff,

you need to be aware, that "revisions" are not meant to be used for =
document versioning. _rev should be called _mvcc_token (or something =
like that) as couchdb only uses this information for concurrency control =
(conflict detection and resolution). you should *never* rely on old, =
non-conflicting versions to be around, as couch does not make any =
guarantees about their availability.

for that reason, I believe that this approach is not working if you seek =
to do a full audit of every change to a database. e.g. when you need to =
compact your database, you might "loose revision" information.


best

Sebastian

On 07.08.2010, at 16:56, Cliff Williams wrote:

> All,
>=20
> Good afternoon.
>=20
> using this url =
http://10.0.0.10:5984/test/_changes?since=3D14966&all_docs=3Dtrue
>=20
> I get the following results
>=20
> {"results":[
> =
{"seq":14971,"id":"ab53d85b92247bcbf251ef9abed4d998","changes":[{"rev":"2-=
b9cfa00cc907767d6bfdeca7dee5008f"}]},
> =
{"seq":14973,"id":"ab53d85b92247bcbf251ef9abed4d512","changes":[{"rev":"3-=
ec350b40276b7c16c09ef485dd8f14aa"}]},
> =
{"seq":14981,"id":"ab53d85b92247bcbf251ef9abed4a63a","changes":[{"rev":"2-=
fe0cb06e7046e9d41d326b3e8c30ed0c"}]},
> =
{"seq":14982,"id":"_design/test","changes":[{"rev":"8-5a565b32bca9b4247954=
3441f23a2e61"}]},
> =
{"seq":14983,"id":"ab53d85b92247bcbf251ef9abed4b4d8","changes":[{"rev":"2-=
ed899403c21b6cca22f7e68f05d92b0c"}]},
> =
{"seq":14984,"id":"ab53d85b92247bcbf251ef9abed47f2f","changes":[{"rev":"2-=
30d5cf6764f2e6c1795d65f272817238"}]},
> =
{"seq":15006,"id":"ab53d85b92247bcbf251ef9abed4e544","changes":[{"rev":"10=
-e6856aac928ff1e2e019b556de8f1760"}]},
> =
{"seq":15007,"id":"ab53d85b92247bcbf251ef9abed49e50","changes":[{"rev":"19=
-d7629eac5a5bd2133a19178fdc9277e6"}]}
> ],
> "last_seq":15007}
>=20
> now if I make an amendment to the document at sequence number 15006 =
and using the same URL I get the following results.
>=20
> {"results":[
> =
{"seq":14971,"id":"ab53d85b92247bcbf251ef9abed4d998","changes":[{"rev":"2-=
b9cfa00cc907767d6bfdeca7dee5008f"}]},
> =
{"seq":14973,"id":"ab53d85b92247bcbf251ef9abed4d512","changes":[{"rev":"3-=
ec350b40276b7c16c09ef485dd8f14aa"}]},
> =
{"seq":14981,"id":"ab53d85b92247bcbf251ef9abed4a63a","changes":[{"rev":"2-=
fe0cb06e7046e9d41d326b3e8c30ed0c"}]},
> =
{"seq":14982,"id":"_design/test","changes":[{"rev":"8-5a565b32bca9b4247954=
3441f23a2e61"}]},
> =
{"seq":14983,"id":"ab53d85b92247bcbf251ef9abed4b4d8","changes":[{"rev":"2-=
ed899403c21b6cca22f7e68f05d92b0c"}]},
> =
{"seq":14984,"id":"ab53d85b92247bcbf251ef9abed47f2f","changes":[{"rev":"2-=
30d5cf6764f2e6c1795d65f272817238"}]},
> =
{"seq":15007,"id":"ab53d85b92247bcbf251ef9abed49e50","changes":[{"rev":"19=
-d7629eac5a5bd2133a19178fdc9277e6"}]},
> =
{"seq":15008,"id":"ab53d85b92247bcbf251ef9abed4e544","changes":[{"rev":"11=
-e01ef09eef56397e345f598fd2069f22"}]}
> ],
> "last_seq":15008}
>=20
>=20
> I think that it is fairly straight forward to ascertain that we are =
looking at latest revision sequences.
>=20
> Is there a way to see all sequence numbers?? I am of course aware of =
the continuous feed but do not wish to use it.
> My use case is for a pharma regulated clinical trials database where i =
need a total audit of every single change made to the database.
>=20
> best regards
>=20
> Cliff