A limitation of ejb or a limitation of jrun?

Question

Hi
I'm working now with jrun designing a few entity beans and i have i problem.
I'm going to explain the situation.
Let's imagine a couple of database tables with the next definition:
TABLE1=(PK(attrib1),....,more attributes.)
TABLE2=(PK(attrib2),....,FK(TABLE1.attrib1),...)
TABLE3=(PK(FK(attrib2),attrib3),........)
At this point i've designed 3 components for representing the elements of the tables
When designing the TABLE3-component i want a method for finding all items at TABLE3 that are related with items at TABLE2 that have a concrete value of attrib1 (FK). it means that i want to return all items of TABLE3 that appear at TABLE2 and with an specific value of the FK.
The syntax of this method should be something like this:
public Collection findByAttrib1(String attrib1) { .. }
Of course this method appears at the Home interface of the TABLE3-component.
When i want to use CMP1.x persistence the app-server refuses to receive a value to be used at the SELECT statement that isn't included at the TABLE3 attributes list. It seems that i only can specify attributes from TABLE3 when i'm going to create the statement at the .xml config file.
So my question is:
is it a limitation of jrun? or this is not allowed at the ejb specification?

Answer 1

When i want to use CMP1.x persistence Why don't you use CMP 2.0 ?
is it a limitation of jrun? Yes, because CMP 1.1 doesn't specify any standard for CMP finder queries at all.
or this is not allowed at the ejb specification?It can be done with CMP 2.0 and EJB QL, like this:
SELECT OBJECT(o) FROM table3 t, IN (t.table2Items) x WHERE x.fk=?1