jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Wolf Benz <euroj...@gmail.com>
Subject Fwd: existing DB: how to begin...
Date Tue, 06 Feb 2007 21:39:02 GMT
Sharing also Nicolas'reply (with his consent)


Yes it's what you'd put into the XSL. You could also write a Java class.

If you want to play with a new technology, you can try XQuery (now in  
Oracle, look at SQL/XML), I think it might do the trick pretty easily.

I haven't done it, but worked a little bit on a freelance project  
with a somewhat similar approach and know well those XML  
vocabularies. I cannot guarantee it's the path to success but you can  
find this out pretty easily and ask for ML support: others might have  

Does it help?


On 2/6/07, Wolf Benz < eurojava@gmail.com> wrote:
Our comp bought this wonderful XXX lock-in system and finally they  
realize it's crap. Yet of course they don't want to throw away all  
historical data. Guess what Dept can solve the Mgr's decisions?  
Right :-)
So now they want US to provide a migration path to solve these  
I see the XML export could very well work, but how best organise  
this? I mean, The node structure I came up with does by far NOT map  
the current schemas.
Guess I'm looking at some heafty XSL transfo work.


On 06 Feb 2007, at 21:24, Nicolas wrote:

OK I get it.

The migration is easier than you think: you can use for instance  
Oracle XML capabilities and generate through a XSL  a XML dump of the  
repository you import (please see system view XML and document view;  
this heavily depends on the volume of your data). You might want to  
ask the mailing list about this use case, but it seems feasible if  
you take certain precaution.

If you have a legacy app accessing this data and you don't want to  
update it, I might have a solution but I am not sure whether this  
would work or not :p

I can elaborate more if you need.


On 2/6/07, Wolf Benz <eurojava@gmail.com > wrote:
The data IS in the repo already (Oracle DB)
Yet now, for the 1st time, I wish to access this data with the JCR  
API. (& adding some stuff like versioning etc)

On 06 Feb 2007, at 21:15, Nicolas wrote:

I don't understand one thing: do you want to migrate your data in a  
repository or use the JCR as an interface?

If you can migrate your data, it's fairly easy: you have to write a  
XML serialization and use importXML (with some caveats). The idea is  
the same as refactoring a DB schema.

Is it clearer?

On 2/6/07, Wolf Benz <eurojava@gmail.com> wrote:
Could you elaborate a bit?

On 06 Feb 2007, at 21:07, Nicolas wrote:


Why not write a transformer to a JCR repository and migrate your old  

my blog! http://www.deviant-abstraction.net !!

On 2/6/07, Wolf Benz < eurojava@gmail.com> wrote:
That's indeed what I meant :-)
The reply's not what I had hoped for, but thx anyway!
If only because this is always the question you get when you try to
pomote "JCR": "can we reuse our old data?" (which is, business wise,
a perfectly valid question as well!!)
--> From yr reply I think the reply to this one is then: yes, but not
without serious (one-time)support of a DBA that does a one-time


On 06 Feb 2007, at 13:23, Jukka Zitting wrote:


On 2/5/07, Wolf Benz < eurojava@gmail.com> wrote:
 > I guess I have a very common question:
 > Is it possible to use JSR-170 when your data ALREADY EXISTS...?
 > How you you map thsi then? ("this" being yr brand new nodes you wish
 > to use in yur new apps & your old DB-data)

You mean you have a custom database schema and want to access it
through the JCR API? Unfortunately I don't know of any JCR
implementations that currently do this.


Jukka Zitting

my blog! http://www.deviant-abstraction.net !!

my blog! http://www.deviant-abstraction.net !!

my blog! http://www.deviant-abstraction.net !!

my blog! http://www.deviant-abstraction.net !!

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message