jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Phukan, Anit" <Anit_Phu...@intuit.com>
Subject RE: recursive XPath in jackrabbit...
Date Wed, 30 Sep 2009 00:01:06 GMT
Thanks Alex!

I was able to use the persistence manager with a local SQL Server DB,
and I could see that some tables were created in the database (residing
in a SQL Server Database Primary Data File on the server data
directory). The tables that were created were the Property and node
tables, with PROP_ID, PROP_DATA as columns for the Property tables, and
NODE_ID, NODE_DATA as columns for the Node table.

I am accessing the nodes directly now to retrieve data and add any
additional properties to my objects.

Thanks a lot for illuminating my thought process on this.

Regards
Anit

-----Original Message-----
From: Alexander Klimetschek [mailto:aklimets@day.com] 
Sent: Tuesday, September 29, 2009 10:16 AM
To: users@jackrabbit.apache.org
Subject: Re: recursive XPath in jackrabbit...

On Tue, Sep 29, 2009 at 18:38, Phukan, Anit <Anit_Phukan@intuit.com>
wrote:
> So is it advisable to use the DBFileBased system wherein we write our
> objects to specific nodes on a DB file system, and add properties
later
> on as opposed to actually creating tables for nodes and properties?

No, the FileSystem is actually a legacy component in the internal
jackrabbit persistence architecture. The important part is the
persistence manager, see [1]. The bundle DB persistence managers are
recommended for performance and stability. They will create their own
tables upon first access in the database, which are not really
intended to be accessed by other applications, as they are internal
and contain most information in binary blobs (node bundles, hence the
name).

The idea behind that is to give you a JCR API, which is completely
independent from whatever storage (eg. relational db is below) and
clients don't have to care about. Unstructured data is supported
without further ado.

How you map your objects onto JCR itself is another topic... you can
use jackrabbit-ocm, jcrom or even drop object mappings and always
access nodes directly (my opinion), as the JCR Node interface is
already kind of a DAO.

[1] http://wiki.apache.org/jackrabbit/PersistenceManagerFAQ

Regards,
Alex

-- 
Alexander Klimetschek
alexander.klimetschek@day.com

Mime
View raw message