Mailing-List: contact users-help@jackrabbit.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@jackrabbit.apache.org
Received-SPF: pass (nike.apache.org: local policy)
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: performance question
Date: Wed, 28 Nov 2007 21:28:53 +0100
Message-ID: <F8E386B54CE3E6408F3A32ABB9A7908A57598E@hai02.hippointern.lan>
In-Reply-To: <510143ac0711281217w538d4a1v6e7ce9d059117282@mail.gmail.com>
Thread-Topic: performance question
Thread-Index: Acgx+752d96pTiJ9SfewfO2nYsX1mwAAI+QQ
References: <20071128175431.GA8046@office.redwerk.com>
 <510143ac0711281217w538d4a1v6e7ce9d059117282@mail.gmail.com>
From: "Ard Schrijvers" <a.schrijvers@hippo.nl>
To: <users@jackrabbit.apache.org>


>=20
> Hi,
>=20
> On Nov 28, 2007 7:54 PM, Eugeny N Dzhurinsky <bofh@redwerk.com> wrote:
> > So in general the entire tree has several domain nodes, and each=20
> > domain node hash set of a nodes which hold the content of domain=20
> > object for given date, and each date node has a lot (up to=20
> several millions) of child nodes.

> Jukka Zitting wrote:
> The large number of child nodes per a "date" node is a clear=20
> warning sign. Accessing a node with a large number (>10k) of=20
> direct child nodes is expensive.
>=20
> Can you rearrange the child nodes to a subtree hierarchy?

I really assumed he meant several millions of child nodes as in several
million descendants...I just cannot believe Eugeny is talking about
millions of child nodes without any hierarchy.=20

Though, an XPath query won't be slower for a flat distribution if I am
correct (actually faster because the hierarchy resolver has an easier
job), but he also seems to suffer from queries that do not perform.=20

Anyway, Eugeny, can you provide us the missing information?

Ard

>=20
> BR,
>=20
> Jukka Zitting
>=20