> 1) I would like to have a Entity Bean that would map not only a record > from my database, but severals, and that I could manipulate that > information in the Entity Bean class to be prepared to handle client > requests. For example, I want create a Entity Bean that represents a > Graph (several records), and I would like that Graph to be initialized > in a specific data structure in my Java class, for I could run a routing > algorithm (Dijkstra). Is this possible?
You can write your own load/save code using BMP (bean managed persistence) - the important thing is to retain the primary key concept. > 2) If I have create a Entity Bean, and if I create Session Beans to make > the callback interface for the clients, I would have several Session > Beans for every client's request. But how many Entity Beans would I have > in my container? Only one, right? All Session Beans would go to the same > Entity Bean? in this context it is not clear what you mean by "Entity Bean". Depending on the container's implementation and configuration you may or may not have more than one instances of the entity bean class. Having said that I think all containers nowdays support a one-instance approach. Access to the entity bean is then synchronized, to get concurrent access to data (only one client requiring a write lock) most containers also support multiple instances... hth dim =========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff EJB-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
