我有很多可能的关系 CohortGroup 和 Employee。每当我将 Employee 插入 CohortGroup hibernate 时，都会从解析表中删除该组并再次插入所有成员，再加上新成员。为什么不直接添加新的？

组内注解:

@ManyToMany(cascade = { PERSIST, MERGE, REFRESH })
@JoinTable(name="MYSITE_RES_COHORT_GROUP_STAFF",
joinColumns={@JoinColumn(name="COHORT_GROUPID")},
inverseJoinColumns={@JoinColumn(name="USERID")})
public List<Employee> getMembers(){
  return members;
}

在Employee的另一边

@ManyToMany(mappedBy="members",cascade = { PERSIST, MERGE, REFRESH } )
public List<CohortGroup> getMemberGroups(){
  return memberGroups;
}

代码片段

Employee emp = edao.findByID(cohortId);
CohortGroup group = cgdao.findByID(Long.decode(groupId));
group.getMembers().add(emp);
cgdao.persist(group);

下面是日志中报的sql

delete from swas.MYSITE_RES_COHORT_GROUP_STAFF where COHORT_GROUPID=?
insert into swas.MYSITE_RES_COHORT_GROUP_STAFF (COHORT_GROUPID, USERID) values (?, ?)
insert into swas.MYSITE_RES_COHORT_GROUP_STAFF (COHORT_GROUPID, USERID) values (?, ?)
insert into swas.MYSITE_RES_COHORT_GROUP_STAFF (COHORT_GROUPID, USERID) values (?, ?)
insert into swas.MYSITE_RES_COHORT_GROUP_STAFF (COHORT_GROUPID, USERID) values (?, ?)
insert into swas.MYSITE_RES_COHORT_GROUP_STAFF (COHORT_GROUPID, USERID) values (?, ?)
insert into swas.MYSITE_RES_COHORT_GROUP_STAFF (COHORT_GROUPID, USERID) values (?, ?)

这种接缝确实效率低下并且会导致一些问题。如果发出多个请求将员工添加到组中，那么有些请求会被覆盖。

像 equals 和 hashCode 这样的接缝可能是造成这种情况的一个原因。下面是这些方法的实现。有任何危险信号吗？

队列组

    @Override
public int hashCode() {
    final int prime = 31;
    int result = getName().hashCode();
    result = prime * result + ((emp == null) ? 0 : emp.hashCode());
    return result;
}
@Override
public boolean equals(Object obj) {
    if (this == obj) {return true;}
    if (!(obj instanceof CohortGroup)) {return false;}
    CohortGroup other = (CohortGroup) obj;
    if(!getName().equals(other.getName())){return false;}
    if (emp == null && other.getOwner() != null) {
        return false;
    } else if (!emp.equals(other.getOwner())) {
        return false;
    }
    return true;
}

员工

       @Override
public boolean equals(Object obj) {
    if (this == obj) {return true;}
    if (obj == null) {return false;}
    if (!(obj instanceof Employee)) {return false;}
    Employee other = (Employee) obj;
    if (EMPLID == null && other.getEMPLID() != null) {
        return false;
    } else if (!EMPLID.equals(other.getEMPLID())) {
        return false;
    }
    return true;
}

   @Override
public int hashCode() {
    final int prime = 31;
    int result = 1;
    result = prime * result + ((EMPLID == null) ? 0 : EMPLID.hashCode());
    return result;
}

我向 CohortGroup 添加了一个 addMember 方法，该方法添加到关系的两侧:

    public void addMember(Employee emp){
    this.getMembers().add(emp);
    emp.getMemberGroups().add(this);

}

继续感谢所有提供帮助的人。

最佳答案

我高度怀疑您没有正确覆盖 equals 和 hashCode。错误地覆盖它们可能会导致您遇到的那种行为(因为散列键用作映射中的键)。仔细检查您对 equals 和 hashCode 做了什么。

使用具有良好equals 和hashCode 的您的 注释实体，此代码(逻辑上等效):

Session session = HibernateUtil.beginTransaction();
Employee emp = (Employee) session.load(Employee.class, 1L);
CohortGroup group = (CohortGroup) session.load(CohortGroup.class, 1L);
group.getMembers().add(emp);
emp.getMemberGroup().add(group); // set the other side too!!
session.saveOrUpdate(group);
HibernateUtil.commitTransaction();

在我的机器上产生以下输出:

08:10:32.426 [main] DEBUG o.h.e.d.AbstractFlushingEventListener - processing flush-time cascades
08:10:32.431 [main] DEBUG o.h.e.d.AbstractFlushingEventListener - dirty checking collections
08:10:32.432 [main] DEBUG org.hibernate.engine.CollectionEntry - Collection dirty: [com.stackoverflow.q2649145.CohortGroup.members#1]
08:10:32.432 [main] DEBUG org.hibernate.engine.CollectionEntry - Collection dirty: [com.stackoverflow.q2649145.Employee.memberGroup#1]
08:10:32.443 [main] DEBUG org.hibernate.engine.Collections - Collection found: [com.stackoverflow.q2649145.CohortGroup.members#1], was: [com.stackoverflow.q2649145.CohortGroup.members#1] (initialized)
08:10:32.448 [main] DEBUG org.hibernate.engine.Collections - Collection found: [com.stackoverflow.q2649145.Employee.memberGroup#1], was: [com.stackoverflow.q2649145.Employee.memberGroup#1] (uninitialized)
08:10:32.460 [main] DEBUG o.h.e.d.AbstractFlushingEventListener - Flushed: 0 insertions, 0 updates, 0 deletions to 2 objects
08:10:32.461 [main] DEBUG o.h.e.d.AbstractFlushingEventListener - Flushed: 0 (re)creations, 2 updates, 0 removals to 2 collections
08:10:32.463 [main] DEBUG org.hibernate.pretty.Printer - listing entities:
08:10:32.473 [main] DEBUG org.hibernate.pretty.Printer - com.stackoverflow.q2649145.CohortGroup{id=1, members=[com.stackoverflow.q2649145.Employee#1]}
08:10:32.474 [main] DEBUG org.hibernate.pretty.Printer - com.stackoverflow.q2649145.Employee{id=1, memberGroup=}
08:10:32.474 [main] DEBUG o.h.p.c.AbstractCollectionPersister - Inserting collection: [com.stackoverflow.q2649145.CohortGroup.members#1]
08:10:32.480 [main] DEBUG org.hibernate.jdbc.AbstractBatcher - about to open PreparedStatement (open PreparedStatements: 0, globally: 0)
08:10:32.491 [main] DEBUG org.hibernate.SQL - insert into MYSITE_RES_COHORT_GROUP_STAFF (COHORT_GROUPID, USERID) values (?, ?)
Hibernate: insert into MYSITE_RES_COHORT_GROUP_STAFF (COHORT_GROUPID, USERID) values (?, ?)
08:10:32.496 [main] TRACE org.hibernate.type.LongType - binding '1' to parameter: 1
08:10:32.497 [main] TRACE org.hibernate.type.LongType - binding '1' to parameter: 2
08:10:32.499 [main] DEBUG o.h.p.c.AbstractCollectionPersister - done inserting collection: 1 rows inserted

No delete before the insert!

By the way, note that you should set the link on both sides when working with bi-directional associations, like I did on the group and on the employee.

Or add defensive link management methods on your classes, for example on CohortGroup:

public void addToMembers(Employee emp) {
    this.getMembers().add(emp);
    emp.getMemberGroup().add(this);
}

public void removeFromMembers(Employee emp) {
    this.getMembers().remove(emp);
    emp.getMemberGroup().remove(this);
}

关于java - hibernate insert to a collection 导致删除，然后再次插入集合中的所有项目，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/2649145/