I whould like to have some suggestions about this situation:
There is a large table, with 60 columns (TABLE_A).
The first 20 columns are always used, while the others are used seldom.
I would like to have 2 classes (LightClass & HeavyClass) to not transfer to client-side all the 60 columns (using LightClass with only the first 20 columns).
Which is the best strategy?
I've tried to create a HeavyClass which inherits from LightClass and I try to use both sub-class and joined-sublclass tab, but both of them are useless, since I do not have different rows for different classes, but the same row can be of one class or another depends on call.
So I've tried to map 20 columns in LightClass and 40 in HeavyClass and map a one-to-one relation between them ... but it do not work for 2 reason:
1) proxy do not work with one-to-one mapping (the second class is always read also if is a proxy) ...
2) and if the 2 tables has the same name there are some exceptions at run-time.

And, ... if possible, I whould not like to have 2 mapping files with 40 columns in common between them.
Thansk for help
Alessandro Rizzi

Take a look at http://www.hibernate.org/41.html

Thanks for your suggestion.
I've seen it, but in this situation I have to copy & paste all the 40 common columns in the mappings. Is there a way to avoid it?
Regards
Alessandro Rizzi

If you use XDoclet you might want to try something like....

AbstractLightClass
... your 20 common fields go here with XDoclet field mappings

Abstract40LessUsedFieldClass extends AbstractLightClass
... your 40 fields go in here with XDoclet field mappings.

LightClass extends AbstractLightClass
Just map the class/table here all field mappings will be inherited
@hibernate.class table="60ColumnTable"

HeavyClass extends Abstract40LessUsedFieldClass
Just map the class/table here all field mappings will be inherited
@hibernate.class table="60ColumnTable"

hope this helped!!!

This could help me, but I do not use XDocLet ... (for many pourpose, not under my control)
Each Hibernate mapping is on a separated file.

That's too bad, it really makes doing your mappings much easier and more intuitive. I'm pretty sure you will have to duplicate your mappings, because even with XDoclet the mapping XML files that are generated would contain duplicated columns. The advantage is that you only have to deal with each column in one place and then let Ant and XDoclet do the dirty work. It would be worth your time (IMO) to discuss this possiblity with whoever is in control. If not I guess you'll just do a simple copy and paste. ;)

good luck!