spring jpa 连表重复 join 如何解决？

Question

UserEntity

@Data
@Accessors(chain = true)
@Entity
@Table(name = "user")
public class UserEntity {
    @Id
    @GeneratedValue(strategy = GenerationType.IDENTITY)
    @Column(name = "id")
    private Long id;

    @OneToOne(mappedBy = "user", fetch = FetchType.LAZY)
    @JsonManagedReference
    private UserArchiveEntity userArchive;
}

UserArchiveEntity

@Data
@Accessors(chain = true)
@Entity
@Table(name = "user_archive")
public class UserArchiveEntity {
    @Id
    @Column(name = "user_id")
    private Long userId = 0L;

    @OneToOne(fetch = FetchType.LAZY)
    @JoinColumn(name = "user_id", insertable = false, updatable = false)
    // 从属关系：从；一定要加这个，避免json序列化循环引用导致报错
    @JsonBackReference
    private UserEntity user;

    @ManyToOne(fetch = FetchType.LAZY)
    @JoinColumn(name = "nation_id", insertable = false, updatable = false)
    private NationEntity nation;

    @ManyToOne(fetch = FetchType.LAZY)
    @JoinColumn(name = "industry_id", insertable = false, updatable = false)
    private IndustryEntity industry;
}

UserRepository

@Repository
public interface UserRepository extends JpaRepository<UserEntity, Long>, JpaSpecificationExecutor<UserEntity> {
    @EntityGraph(attributePaths = {"userArchive", "userArchive.nation", "userArchive.industry"})
    @NotNull
    @Override
    Page<UserEntity> findAll(@NotNull Specification<UserEntity> specification, @NotNull Pageable pageable);

我的查询如下：

public List<UserEntity> recommend(int size) throws Exception {
        UserEntity userEntity = UserContext.get();
        Optional<UserArchiveEntity> optionalUserArchiveEntity = this.userArchiveRepository.findByUserId(userEntity.getId());
        if (optionalUserArchiveEntity.isEmpty()) {
            throw new Exception("用户档案记录不存在");
        }
        UserArchiveEntity userArchiveEntity = optionalUserArchiveEntity.get();
        Integer gender = userArchiveEntity.getGender();
        Integer judgeGender = gender == UserArchiveEntity.GENDER_MALE ? UserArchiveEntity.GENDER_FEMALE : UserArchiveEntity.GENDER_MALE;
        Specification<UserEntity> specification = (root, query, criteriaBuilder) -> {
            Predicate predicate = criteriaBuilder.conjunction();
            Join<UserEntity, UserArchiveEntity> userArchiveEntityJoin = root.join("userArchive");
            return predicate;
        };
        Sort sort = Sort.by(Sort.Direction.DESC, "infoCompletedRatio");
        Pageable pageable = PageRequest.of(0, size, sort);
        return this.userRepository.findAll(specification, pageable).getContent();

    }

出现的问题：

1. 上述代码生成的sql针对 userArchive 会 left join 两次！即 user left join userarchive on xxx left join userarchive on xxx left join nation on xx on left join industry on xx，该如何避免重复 join
2. findAll 方法针对不同的业务场景关联的实体要求不一样；但是 spring jpa 没法声明方法签名一致的方法，也没法随意声明不同名称的方法（需要符合 spring jpa 的规范名称才会被正确解析），且由于要支持动态条件查询，必须要有 specification 这种 criterial 标准api 。该如何针对不同业务场景实现不同的模型关联？

无

Answer 1

• 第一个问题join userArchive重复，是因为实体关联 @OneToOne 是左连接，在Specification 中手动 join 用的是内连接，所以被判定为不同的连接方式导致。在specification 中指定 root.join("userArchive", JoinType.LEFT) 指明为左连接，就不会出现重复连接问题。
• 第二个问题，目前没想到好的方法，等待高手回答。

Answer 2

使用QueryDSL查询框架来解决
链接：博客园上关于QueryDSL的介绍和使用说明：为 JPA 插上翅膀的 QueryDSL

或者在字段上添加注解

@JoinColumn(name = "column_name", foreignKey = @ForeignKey(value = ConstraintMode.NO_CONSTRAINT))

使得原来的外键链接变成逻辑链接，在查询时就会分开查