JPA中Address共享parent_id关联Person与Company的问题解决问询

方案一：继承映射（优雅且符合JPA规范） #

核心思路是将Person和Company抽象为同一个父实体，让Address关联这个父实体，从而实现通过单一parent_id列关联两种实体，同时避免冗余字段。

1. 创建抽象父实体Parent #

@Entity
@Inheritance(strategy = InheritanceType.JOINED) // 采用Joined继承，避免单表冗余
@DiscriminatorColumn(name = "entity_type", discriminatorType = DiscriminatorType.STRING)
public abstract class Parent {
    @Id
    private String id;

    @OneToOne(cascade = CascadeType.ALL, orphanRemoval = true, mappedBy = "parent")
    private Address address;

    // Getter & Setter
}

2. 修改Person实体继承Parent #

通过@AttributeOverride保留原表的主键字段名（person_id）：

@Entity
@DiscriminatorValue("PERSON")
@AttributeOverride(name = "id", column = @Column(name = "person_id"))
public class Person extends Parent {
    // 这里添加Person专属字段，无需重复定义personId
}

3. 修改Company实体继承Parent #

同样保留原主键字段名（company_id）：

@Entity
@DiscriminatorValue("COMPANY")
@AttributeOverride(name = "id", column = @Column(name = "company_id"))
public class Company extends Parent {
    // 这里添加Company专属字段，无需重复定义companyId
}

4. 调整Address实体关联Parent #

@Entity
public class Address {
    @Id
    private Long id;

    @OneToOne
    @JoinColumn(name = "parent_id") // 唯一外键列，关联Parent的ID
    private Parent parent;

    @OneToMany(mappedBy = "address")
    private List<Area> areas = new ArrayList<>();

    // Getter & Setter
}

5. 保持Area实体不变 #

@Entity
public class Area {
    @Id
    private Long id;

    @ManyToOne
    @JoinColumn(name = "address_id")
    private Address address;

    // Getter & Setter
}

方案优势 #

• 完全符合JPA规范，Hibernate可以自动维护parent_id字段，无需手动干预
• 查询Person/Company时，直接通过父实体的address关联即可获取地址，支持懒加载或预加载
• 无冗余字段，仅用parent_id实现多实体关联

方案二：手动维护外键（无需继承） #

如果不想引入继承关系，可以通过手动维护parent_id字段，同时将Address中关联Person/Company的字段设为只读，避免Hibernate的冲突检测。

1. 修改Address实体 #

添加parentId字段用于手动维护，同时将person/company设为只读并添加过滤条件：

@Entity
public class Address {
    @Id
    private Long id;

    private String parentId; // 手动维护的外键字段

    @OneToOne
    @JoinColumn(name = "parent_id", insertable = false, updatable = false)
    @Where(clause = "exists (select 1 from person p where p.person_id = parent_id)")
    private Person person;

    @OneToOne
    @JoinColumn(name = "parent_id", insertable = false, updatable = false)
    @Where(clause = "exists (select 1 from company c where c.company_id = parent_id)")
    private Company company;

    @OneToMany(mappedBy = "address")
    private List<Area> areas = new ArrayList<>();

    // Getter & Setter
}

2. 修改Person实体 #

通过@PrePersist和@PreUpdate钩子，在保存/更新时自动设置Address的parentId：

@Entity
public class Person {
    @Id
    private String personId;

    @OneToOne(cascade = CascadeType.ALL, orphanRemoval = true)
    private Address address;

    @PrePersist
    @PreUpdate
    private void syncAddressParentId() {
        if (address != null) {
            address.setParentId(this.personId);
        }
    }

    // Getter & Setter
}

3. 修改Company实体 #

同样添加钩子自动维护Address的parentId：

@Entity
public class Company {
    @Id
    private String companyId;

    @OneToOne(cascade = CascadeType.ALL, orphanRemoval = true)
    private Address address;

    @PrePersist
    @PreUpdate
    private void syncAddressParentId() {
        if (address != null) {
            address.setParentId(this.companyId);
        }
    }

    // Getter & Setter
}

查询时关联获取Address #

无论采用哪种方案，都可以通过以下方式实现查询Person/Company时一并加载Address：

方式1：使用@EntityGraph（推荐） #

在Repository方法上添加注解，指定预加载关联字段：

public interface PersonRepository extends JpaRepository<Person, String> {
    @EntityGraph(attributePaths = "address")
    Optional<Person> findByPersonId(String personId);
}

public interface CompanyRepository extends JpaRepository<Company, String> {
    @EntityGraph(attributePaths = "address")
    Optional<Company> findByCompanyId(String companyId);
}

方式2：使用JPQL的JOIN FETCH #

public interface PersonRepository extends JpaRepository<Person, String> {
    @Query("SELECT p FROM Person p JOIN FETCH p.address WHERE p.personId = :id")
    Optional<Person> findByIdWithAddress(@Param("id") String id);
}

内容的提问来源于stack exchange，提问作者Nagendra