在JPA中选择具有子实体的许多实体的最佳优化方法是什么？

Question

Let's say we have:

@Entity public class Order {
  @Id private int id;
  @OneToMany(mappedBy="order") private List<Item> items;
  ...
}

and

@Entity public class Item {
  @Id private int id;
  @ManyToOne private Order order;
  ...
}

And let's say there is 10.000 orders with each having 20 items.

We need to iterate thought all order and all their items. What is the best way to do it in JPA?

My issue is, if I just iterate the elements like:

for (Order order: em.createTypeQuery("select o from Order o", Order.class).getResultList()) {
    report.printOrder(order);
    for (Item item: order.getItems()) {
        report.printOrderItem(item);
    }
}

it will result in 10.001 sql queries: 1 time: select * from order 10.000 times: select * from item where order_id = ?

Is there any way to optimize it? Two queries? One query?

(We are using EclipseLink)

Thanks.

Answer 1

You may also want to consider the EclipseLink query Hint "eclipselink.batch" with value "o.items". This results in two queries but can be more effecient than once large joined query.

Answer 2

您可以使用join fetch （这也需要distinct因为join fetch具有join语义）：

select distinct o from Order o join fetch o.items