Spring with JPA and Hibernate

Spring integrates with JPA and Hibernate through a set of runtime mechanisms that manage data access, persistence, and transaction boundaries for you. When you use Spring Data JPA or configure JPA in a Spring application, Spring automatically creates and manages an EntityManager for each transaction. This is achieved by proxying the EntityManager so that you can inject and use it anywhere in your application without worrying about its lifecycle.

The EntityManager is bound to the current transaction context. When a transaction begins, Spring obtains an EntityManager from the configured EntityManagerFactory and binds it to the current thread. This ensures that all data access operations within the transaction use the same persistence context.

Under the hood, Hibernate acts as the default JPA implementation. It manages a Session for each EntityManager. This Session is tightly coupled with a JDBC connection, which is also bound to the transaction. When you perform operations like persist, merge, or remove, Hibernate tracks changes in the session and prepares SQL statements.

On a flush operation, Hibernate synchronizes the in-memory state with the database by executing the necessary SQL statements using the JDBC connection. When the transaction commits, Hibernate flushes any remaining changes, commits the JDBC transaction, and releases the connection. If a rollback occurs, Hibernate discards uncommitted changes and cleans up resources accordingly.

This internal orchestration ensures consistent data access, efficient resource management, and reliable transaction handling across your Spring application.