JPA Persistence Context Explained

Direct Answer

A persistence context is the set of entity instances managed by one EntityManager. It guarantees one managed instance per entity identity, tracks changes for flush, and serves as the first-level cache. In Spring it is commonly transaction-scoped, but JPA also supports extended persistence contexts, so it is not inherently identical to a transaction.

Code Example

@Service
public class OrderService {
    private final OrderRepository orderRepository;
    OrderService(OrderRepository r) { this.orderRepository = r; }

    @Transactional
    public void demo(Long orderId) {
        var first = orderRepository.findById(orderId).orElseThrow(); // SELECT
        var second = orderRepository.findById(orderId).orElseThrow(); // no SELECT (cached)
        System.out.println(first == second); // true — same managed object

        first.setStatus("SHIPPED"); // dirty checking → UPDATE at commit
    }
}

Common Mistakes & Pitfalls

• Confusing first- and second-level cache — the persistence context is the first-level cache (per transaction, always on); the second-level cache is optional and shared across transactions.
• Calling save() after every change — redundant for managed entities; dirty checking handles it.
• Mutating detached entities — outside @Transactional, entities are detached and changes won't sync.
• LazyInitializationException — accessing a lazy relation after the context closes; load it inside the transaction or use a fetch join.