POJOs and createNativeQuery

Question

We are having lots of complex sql and tonnes of tables in our schema and intend to use JPA(present system is on JDBC). We intend to lower the development time by using the native query support in JPA(createNativeQuery) and reuse the existing SQLs ; so that it plays well with UI which is on ADF.
So, I have a native query which selects some columns; and i also have a suitable entity(POJO) which maps to the columns that are fetched in the query. Note that the query fetches columns from multiple tables and hence the POJO resembles something like a value object (though in the present parlance its called as an Enitity). All is fine till here.
Now i need to group by certain columns; how should the result list be now accessed - do i need to create another entity(POJO) which maps to this structure of groupby column name(s) and its respective counts? Is this a fine(and elegant) solution; or can you think of anything else?

Answer 1

It depends,
SqlResultSetMapping is cool. But, I think not in your case.
If you map the values to an entity, the returned entities become managed by the persistence context.
An entity is a bit more than a single pojo, it must be defined as an entity, and must have an id.
So you must create entities for all results of your queries, so all result must have an id.
This is quite limited in JPA.
so maybe single transformator classes could be better choice.
if you can use hibernate instead of "standard jpa", you can use resulttranformator, it would be more useful in your case :)