Do you need a database transaction for reading data?

All database statements are executed within the context of a physical transaction, even when we don’t explicitly declare transaction boundaries (BEGIN/COMMIT/ROLLBACK).

If you don't declare transaction boundaries, then each statement will have to be executed in a separate transaction (autocommit mode). This may even lead to opening and closing one connection per statement unless your environment can deal with connection-per-thread binding.

Declaring a service as @Transactional will give you one connection for the whole transaction duration, and all statements will use that single isolation connection. This is way better than not using explicit transactions in the first place.

On large applications, you may have many concurrent requests, and reducing database connection acquisition request rate will definitely improve your overall application performance.

JPA doesn't enforce transactions on read operations. Only writes end up throwing a TransactionRequiredException in case you forget to start a transactional context. Nevertheless, it's always better to declare transaction boundaries even for read-only transactions (in Spring @Transactional allows you to mark read-only transactions, which has a great performance benefit).

Now, if you use declarative transaction boundaries (e.g. @Transactional), you need to make sure that the database connection acquisition is delayed until there is a JDBC statement to be executed. In JTA, this is the default behavior. When using RESOURCE_LOCAL, you need to set the hibernate.connection.provider_disables_autocommit configuration property and make sure that the underlying connection pool is set to disable the auto-commit mode.

Accoring to my experience with JPA implementation in J2EE, a Transaction manager is always needed in order to perform CRUD operation safety, by guaranteeing a rollback in order to preserve data integrity.

Enterprise applications use different resources to save data and send messages like a database or message queue. If we want to query these resources sequentially and to cancel the whole operation once a problem occurs, we have to put this query in a unit of work so that will be executed as a whole.

You could define it:

• by using related annotations (as shown in the questions); in this way, the container loads automatically the transaction manager for a given persistence context;

• by injecting manually the transaction manager, as follows:

  public class sample {
  
      @PersistenceContext
      EntityManager em;
  
      // Injected transaction manager
      @Inject
      UserTransaction utx;
  
      private static final String[] GAME_TITLES = {
          "Super Mario Brothers",
          "Mario Kart",
          "F-Zero"
      };
  
      private void clearData() throws Exception {
          utx.begin();
          em.joinTransaction();
          System.out.println("Dumping old records...");
          em.createQuery("delete from Game").executeUpdate();
          utx.commit();
      }
  
      private void insertData() throws Exception {
          utx.begin();
          em.joinTransaction();
          System.out.println("Inserting records...");
          for (String title : GAME_TITLES) {
              Game game = new Game(title);
              em.persist(game);
          }
          utx.commit();
          // clear the persistence context (first-level cache)
          em.clear();
      }
  
      // ...
  
  }
  

Spring Data, as JPA-spec implementation, may follow the same approach.

You could find more information by reading the following article: Java_Persistence/Transactions.