hi marcus,
[EMAIL PROTECTED] wrote:
Hi,
just looking for some advice regarding the design of a database application that is required associate archives of XML datafiles with more conventional relational type data. Each XML file will be approximately 10-50K bytes and the archive rate will be more than 50,000 files / year. I would like to use Ojb as the mapping layer.
Is there a standard practice for handling XML files in association with Ojb. Should they be stored within the RDBMS - or just as disk files with a look up key?
there is no general rule. it depends most on your requirements.
I see 3 options:
1. store documents as flat files and keep only file names in the db.
2. store documents in CLOB columns in the db
3. parse xml documents to a DOM-tree and store the DOM-tree in the db. (this is the persistent DOM approach)
1. is the most simple approach.
2. ha sthe advantage that documents and pointers to documents are kept together in the db. that makes management easy.
3. can have advantages if you are more interested in working with DOM representations of your XML documents. you can also avoid reparsing documents after retrieving them from the db.
the downside is that you need a erm model of xml DOM.
you could start with http://dbdom.sourceforge.net/ for such a model.
With the ojb reversedb tool you can then generate the repository.xml mapping file...
I have read that there is Ojb activity to develop support layers for XML-based databases? If so would Apache Xindice be a suitable candidate for an XML database?
this is currently only an idea. Xindice will be the natural candidate. But afaik currently nobody is working on this.
cheers, Thomas
Thanks in advance,
Marcus
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
