Question

你可以解释一下@JoinColumn注释在没有任何其他关系注释的情况下对一个字段的默认行为。

@Entity
class Employee{
  @Id int id;

  @JoinColumn(name = "man_id")
  Employee manager;

}

jpa是如何进行的？如果员工是与部门有关的双向关系的多方面，其中fetchtype lazy指定在双方，当我访问（询问大小的集合）getEmployees部门实体的属性时，eclipcelink生成sql并从员工实体中选择所有记录。

SELECT ID FROM deps WHERE (ID = ?)
SELECT ID, name, man_id, dep_id FROM emps WHERE (dep_id = ?)
additional requests
SELECT ID, name, man_id, dep_id FROM emps WHERE (ID = ?)
SELECT ID, name, man_id, dep_id FROM emps WHERE (ID = ?)
SELECT ID, name, man_id, dep_id FROM emps WHERE (ID = ?)
.................

如果我用@OnetoOne（实际上是我的意思）将这个属性标记为关系并设置fetchtype为lazy，那么＆＃34;的工作正确＆＃34;这意味着在我向员工询问之前，不会提取经理财产。

SELECT ID FROM deps WHERE (ID = ?)
SELECT ID, name, man_id, dep_id FROM deps WHERE (dep_id = ?)

Answer 1

By not putting a relationship type/annotation, actually means JPA would not treat your attribute as an entity.

Perhaps, you already had your schema in place that's why your queries are working. However, I experimented on this and started with a fresh database. The behavior is different from what you expected. With Hibernate, my chosen JPA provider (with hbm2ddl property set to create), it generated the ff. schema:

create table Employee (
    id bigint not null auto_increment,
    manager tinyblob,
    name varchar(255),
    primary key (id)
)

Since my manager is an Employee class implementing Serializable interface, it was by default mapped to a BLOB type column.

Also by not mapping it with any relationship annotations, means that there will be no foreign key constraints applied to your manager column.

JoinColumn默认行为

1 个答案: