Hello Jean-Baptiste, It is excellent for you to take time to pen down your appreciation for OpenJPA. It was a pleasure to see that you found OpenJPA FetchPlan a powerful feature. The good news is you are not alone. During the session on JPA during JavaOne 2010, th expert panel members invited requirements from the audience and FetchPlan featured as one of the top priority items. FetchPlan would be a game changer for many other JPA providers. These providers often use a fixed SQL for access because they lack the capability to fetch a customizable closure as OpenJPA can. The misplaced FetchType.LAZY and EAGER (misplaced because sub-graph specification does not belong to type level annotations) helps these providers to generate a SQL which is somewhat more sophisticated than a flat ResultSet -- but the content is invariant for a type nevertheless in these implementations. This shortcut of fixed SQL based design helped these providers in the past significantly in performance as compared to OpenJPA where ever SQL was regenerated because FetchPlan can vary dynamically at per-request basis. However, with introduction of Prepared Query Caches in OpenJPA since v1.2 had closed the performance gap because now OpenJPA can also reuse SQL as long as the fetch plan remains unchanged. Thank you again for your enthusiastic support of OpenJPA and wish you build great systems on the rock solid foundation of a well-designed persistent layer.
Regards -- Pinaki ----- Pinaki -- View this message in context: http://openjpa.208410.n2.nabble.com/OpenJPA-rock-and-fetch-plan-is-a-killing-feature-tp5978322p5980245.html Sent from the OpenJPA Users mailing list archive at Nabble.com.
