Need advice on xml data / database

[Castor Castro]

Hi folks,

I am currently reviewing for the SCWCD. Part of this quest is to develop a personal project which should apply all I learned. I am looking on creating a web based Leaves Tracking/Filing system which I can offer in our office as well. This system make filing sick/vacation leaves paperless.

As to the subject of this post, I am going to design my database. And here is how I am planning to do it:

Data will be stored in a CLOB field as an xml data. Data is identified by an id of course (say employee id). I will be using Castor to marshall/unmarshall the data. Now, what if a change in field happens? Let's say, another field needs to be added. My solution is that each Castor class files should be serialized and stored in the db as well and identified by a version number. The xml data should also have a version number to reference the Castor class. Upon startup of the application, as part of initialization, those classes will be deserialized in a directory identified by its package. And the story goes on...

Well, I guess that's all I wanted to say. I would like to ask expert advice from you guys. Whether to proceed or not and why. Or you could also give some advice on how I should do it.

Thanks