Hibernate save vs persist vs saveOrUpdate

Direct Answer

EntityManager.persist is the standard JPA operation for making a new entity managed. Hibernate's Session.save, update, and saveOrUpdate are older native APIs and are deprecated in modern Hibernate; new code normally uses persist for new entities and merge for detached state. Spring Data repository.save chooses persist or merge from its new-entity detection.

Code Example

import jakarta.persistence.*;

// JPA standard: persist() inserts a transient entity, returns void
@PersistenceContext EntityManager em;
void createWithJpa(Order order) {
    em.persist(order);            // order must have NO id yet
}

// Hibernate native: save() returns the generated id immediately
// Long id = (Long) session.save(order);

// Modern Spring Data JPA: one method handles insert AND update
interface OrderRepository extends JpaRepository<Order, Long> {}

class OrderService {
    private final OrderRepository repo;
    OrderService(OrderRepository repo) { this.repo = repo; }

    Order saveOrder(Order order) {
        // null id  -> SimpleJpaRepository calls persist() (INSERT)
        // set id   -> it calls merge()  (SELECT then INSERT/UPDATE)
        return repo.save(order);
    }
}

Common Mistakes & Pitfalls

• Using Hibernate Session instead of EntityManager — calling session.save() in Spring services ties code to Hibernate; prefer repositories or EntityManager.
• Calling persist() on a detached entity — persist() only works on transient entities and throws otherwise; use merge() or repository.save().
• Assuming save() always INSERTs — with a non-null ID, Spring Data calls merge(), which does a SELECT first; if it's truly new, keep the ID null to avoid the extra query.