java - JPA 2 实体中的多个 @CollectionTable 导致不相关的外部联接

Question

您好，我有以下 JPA 实体类。它有两个用于组和联系人 ID 的 @CollectionTable 映射。我很感兴趣从 contact_details 表中获取唯一联系人 ID 的列表，并在下面的 Users 类变量 contacts 中引用它们:

@Entity
@Table(name = "users")
public class User {

@Id
@GeneratedValue(generator="uuid")
@GenericGenerator(name = "uuid", strategy = "uuid2")
private UUID id;

@NotBlank
private String username;

@NotBlank
@Column(name = "name")
private String name;

@ElementCollection(fetch = FetchType.EAGER)
@CollectionTable(name = "users_groups", joinColumns = @JoinColumn(name = "user_id"))
@Column(name = "[group]")
private List<String> groups = new ArrayList<>();

@ElementCollection(fetch = FetchType.EAGER)
@CollectionTable(name = "contact_detail", joinColumns = @JoinColumn(name = "user_id"))
@Column(name = "[id]")
private List<String> contacts = new ArrayList<String>();
}

当我从数据库获取用户时，它会在 hibernate 日志中生成以下查询

select user0_.id, user0_.password, user0_.username, contacts1_.user_id as user_id8_4_1_, contacts1_."id" as id1_2_1_, groups2_.user_id as user_id1_4_2_, groups2_."group" as group2_5_2_ 
from users user0_ 
left outer join contact_detail contacts1_ on user0_.id=contacts1_.user_id 
left outer join users_groups groups2_ on user0_.id=groups2_.user_id 
where user0_.id=?

由于 contact_detail 和 users_groups 上的左外连接，它实际上会多次检索到相同的 contact_id。这是具有多个“47e5b98a-2116-4ad9-b773-3acc99e2c83c”联系人 ID 的实体的 JSON 表示形式

{
user: {
    id: "d3b3be2a-8a2a-48ac-94dd-fd85faf1a8ff",
    username: "shiv",
    firstName: "Shivam",
    lastName: "Sinha",
    groups: [
        "ADMIN",
        "USER",
        "ROOT"
    ],
    expired: false,
    locked: false,
    credentialsExpired: false,
    enable: true,
    birthday: "2015-05-18",
    joined: "2015-05-18",
    gender: "M",
    contactDetails: null,
    contacts: [
        "47e5b98a-2116-4ad9-b773-3acc99e2c83c",
        "47e5b98a-2116-4ad9-b773-3acc99e2c83c",
        "47e5b98a-2116-4ad9-b773-3acc99e2c83c"
    ]
}
}

但是，当我完全删除以下组类变量时:

@ElementCollection(fetch = FetchType.EAGER)
@CollectionTable(name = "users_groups", joinColumns = @JoinColumn(name = "user_id"))
@Column(name = "[group]")
private List<String> groups = new ArrayList<>();

并尝试从数据库中检索用户。 Hibernate生成以下sql:

select user0_.id, user0_.password, user0_.username, contacts1_.user_id as     user_id8_4_1_, contacts1_."id" as id1_2_1_ 
from users user0_ left outer join contact_detail contacts1_ on user0_.id=contacts1_.user_id 
where user0_.id=?

这是实体的 JSON 表示形式:

{
user: {
    id: "d3b3be2a-8a2a-48ac-94dd-fd85faf1a8ff",
    username: "shiv",
    firstName: "Shivam",
    lastName: "Sinha",
    expired: false,
    locked: false,
    credentialsExpired: false,
    enable: true,
    birthday: "2015-05-18",
    joined: "2015-05-18",
    gender: "M",
    contacts: [
        "47e5b98a-2116-4ad9-b773-3acc99e2c83c"
    ]
}
}

它仅包含唯一的联系人 ID“47e5b98a-2116-4ad9-b773-3acc99e2c83c”。这正是我们所期望的。

所以我的问题是如何实现同样的事情，而不必删除 groups 类变量或从 List<String> contacts 更改数据类型至Set<String> contacts

Answer 1

现在发生的行为正如从 hibernate 中看到的那样 FAQ .这里的常见问题解答中也提到了解决方案

如果您想通过 JPA 执行此操作，请参阅 here .

以下是使用以下规范在 Spring 数据中的实现

 public class UserSpecifications {

    public static Specification<User> findUniqueUser(Integer userId){

         return new Specification<User>() {

            @Override
            public Predicate toPredicate(Root<User> userRoot,
                    CriteriaQuery<?> criteriaQ, CriteriaBuilder critieriaB) {
                criteriaQ.distinct(true);
                return critieriaB.equal(userRoot.get(User_.id), 1);
            }

         };
    }
}

查找记录

User user = userRepository.findOne(UserSpecifications.findUniqueUser(1));