Thanks Steve! Let me tell you my venture into mapping.xml.
This is the product I am working for: diff. modules of this product exchange information by passing raw xml over http (legacy stuff, no open standards followed). I am trying to bring in Castor for data binding:-) The xml structure for any object "A" looks somewhat like this: <?xml version='1.0' encoding='UTF-8' ?> <!DOCTYPE blah SYSTEM 'http://10.72.181.103:8088/server.dtd'> <blah version='1.0' xmlns='http://www.blah.com/xml'> <results passes="true"> <A> <instanceOfA> </instanceOfA> <instanceOfA> </instanceOfA> .... </A> </results> </blah> Now this is the problem. I have to write schemas for 100s of different objects like A. But then all the schemas will begin with blah and results and then the actual "meat". So, I decided to go for mapping (and thanks to its location attribute) like the below: <mapping> <class name="CollectionOfA" auto-complete="true"> <map-to xml="blah" /> <field name="listOfA" collection="arraylist" type="BeanOfA" > <bind-xml name="instanceOfA" location="results/A" /> </field> </class> <class name="BeanOfA" auto-complete="true"> .....and all the properties of A goes here </class> </mapping> Do you see any solution for defining schemas for all those beans where every xml has those blah/results to begin with? ~Sandeep On 1/9/07, Stephen Winnall <[EMAIL PROTECTED]> wrote:
I think that working from an XML schema is to be preferred if you have that option. The advantages are 1) you can do a good design for your XML; 2) you can reuse any standards which might be available. Since data tends to hang around much longer than programs (and get processed by more than one language), I'd spend more time making sure the data was as well organised as possible. Steve On 9 Jan 2007, at 10:33, Sandeep wrote: > Hi > > Which is better: > 1. Define a schema for a java bean, auto-generate the source code and > get the validation done by castor while populating the bean from a xml > instance. > > 2. Hand code the java bean. Define a mapping file. Populate the java > bean from a xml instance through castor. I don't think I need to > define a schema here (correct me, if I am wrong) > > Advantage for 1) is validation of data types and values can be > performed by castor. > Advantages for 2) are flexibility to add application specific helper > methods to the bean, or when you cannot touch the source for the > bean. > > Obviously, advantage of one is the disadvantage for the other. > > So, how can I get the best of both worlds? I want to hand code my bean > class (not use SourceCodegenerator). I want to define the schema apart > from the mapping so that castor can validate the xml instance against > the schema for both data types and some values. > > > TIA > > ~Sandeep > > --------------------------------------------------------------------- > To unsubscribe from this list please visit: > > http://xircles.codehaus.org/manage_email > --------------------------------------------------------------------- To unsubscribe from this list please visit: http://xircles.codehaus.org/manage_email
--------------------------------------------------------------------- To unsubscribe from this list please visit: http://xircles.codehaus.org/manage_email
