【106期】面试官问:Java 多线程如何实现批量拆分 List 导入数据库?

一、前言

前两天做了一个导入的功能,导入开始的时候非常慢,导入2w条数据要1分多钟,后来一点一点的优化,从直接把list怼进Mysql中,到分配把list导入Mysql中,到多线程把list导入Mysql中。时间是一点一点的变少了。非常的爽,最后变成了10s以内。下面就展示一下过程。

二、直接把list怼进Mysql

使用mybatis的批量导入操作:

 @Transactional(rollbackFor = Exception.class)
    public int addFreshStudentsNew2(List<FreshStudentAndStudentModel> list, String schoolNo) {
if (list == null || list.isEmpty()) {
    return 0;
}
List<StudentEntity> studentEntityList = new LinkedList<>();
List<EnrollStudentEntity> enrollStudentEntityList = new LinkedList<>();
List<AllusersEntity> allusersEntityList = new LinkedList<>();

for (FreshStudentAndStudentModel freshStudentAndStudentModel : list) {

    EnrollStudentEntity enrollStudentEntity = new EnrollStudentEntity();
    StudentEntity studentEntity = new StudentEntity();
    BeanUtils.copyProperties(freshStudentAndStudentModel, studentEntity);
    BeanUtils.copyProperties(freshStudentAndStudentModel, enrollStudentEntity);
    String operator = TenancyContext.UserID.get();
    String studentId = BaseUuidUtils.base58Uuid();
    enrollStudentEntity.setId(BaseUuidUtils.base58Uuid());
    enrollStudentEntity.setStudentId(studentId);
    enrollStudentEntity.setIdentityCardId(freshStudentAndStudentModel.getIdCard());
    enrollStudentEntity.setOperator(operator);
    studentEntity.setId(studentId);
    studentEntity.setIdentityCardId(freshStudentAndStudentModel.getIdCard());
    studentEntity.setOperator(operator);
    studentEntityList.add(studentEntity);
    enrollStudentEntityList.add(enrollStudentEntity);

    AllusersEntity allusersEntity = new AllusersEntity();
    allusersEntity.setId(enrollStudentEntity.getId());
    allusersEntity.setUserCode(enrollStudentEntity.getNemtCode());
    allusersEntity.setUserName(enrollStudentEntity.getName());
    allusersEntity.setSchoolNo(schoolNo);
    allusersEntity.setTelNum(enrollStudentEntity.getTelNum());
    allusersEntity.setPassword(enrollStudentEntity.getNemtCode());  //密码设置为考生号
    allusersEntityList.add(allusersEntity);
}
    enResult = enrollStudentDao.insertAll(enrollStudentEntityList);
    stuResult = studentDao.insertAll(studentEntityList);
    allResult = allusersFacade.insertUserList(allusersEntityList);

if (enResult > 0 && stuResult > 0 && allResult) {
    return 10;
}
return -10;
    }

Mapper.xml

  <insert id="insertAll" parameterType="com.dmsdbj.itoo.basicInfo.entity.EnrollStudentEntity">
insert into tb_enroll_student
<trim prefix="(" suffix=")" suffixOverrides=",">
id,  
remark,  
nEMT_aspiration,  
nEMT_code,  
nEMT_score,  
student_id,  
identity_card_id,  
level,  
major,  
name,  
nation,  
secondary_college,  
operator,  
sex,  
is_delete,  
account_address,  
native_place,  
original_place,  
used_name,  
pictrue,  
join_party_date,  
political_status,  
tel_num,  
is_registry,  
graduate_school,  
create_time,  
update_time</trim>
values
<foreach collection="list" item="item" index="index" separator=",">
(
#{item.id,jdbcType=VARCHAR},
#{item.remark,jdbcType=VARCHAR},
#{item.nemtAspiration,jdbcType=VARCHAR},
#{item.nemtCode,jdbcType=VARCHAR},
#{item.nemtScore,jdbcType=VARCHAR},
#{item.studentId,jdbcType=VARCHAR},
#{item.identityCardId,jdbcType=VARCHAR},
#{item.level,jdbcType=VARCHAR},
#{item.major,jdbcType=VARCHAR},
#{item.name,jdbcType=VARCHAR},
#{item.nation,jdbcType=VARCHAR},
#{item.secondaryCollege,jdbcType=VARCHAR},
#{item.operator,jdbcType=VARCHAR},
#{item.sex,jdbcType=VARCHAR},
0,
#{item.accountAddress,jdbcType=VARCHAR},
#{item.nativePlace,jdbcType=VARCHAR},
#{item.originalPlace,jdbcType=VARCHAR},
#{item.usedName,jdbcType=VARCHAR},
#{item.pictrue,jdbcType=VARCHAR},
#{item.joinPartyDate,jdbcType=VARCHAR},
#{item.politicalStatus,jdbcType=VARCHAR},
#{item.telNum,jdbcType=VARCHAR},
#{item.isRegistry,jdbcType=TINYINT},
#{item.graduateSchool,jdbcType=VARCHAR},
now(),
now()
)   
</foreach>
  </insert>

代码说明:

底层的mapper是通过逆向工程来生成的,批量插入如下,是拼接成类似:insert into tb_enroll_student()values (),()…….();

这样的缺点是,数据库一般有一个默认的设置,就是每次sql操作的数据不能超过4M。这样插入,数据多的时候,数据库会报错Packet for query is too large (6071393 > 4194304). You can change this value on the server by setting the max_allowed_packet' variable.,虽然我们可以通过

类似 修改 my.ini 加上 max_allowed_packet =6710886467108864=64M,默认大小4194304 也就是4M

修改完成之后要重启mysql服务,如果通过命令行修改就不用重启mysql服务。

完成本次操作,但是我们不能保证项目单次最大的大小是多少,这样是有弊端的。所以可以考虑进行分组导入。

三、分组把list导入Mysql中

同样适用mybatis批量插入,区别是对每次的导入进行分组计算,然后分多次进行导入:

 @Transactional(rollbackFor = Exception.class)
    public int addFreshStudentsNew2(List<FreshStudentAndStudentModel> list, String schoolNo) {
if (list == null || list.isEmpty()) {
    return 0;
}
List<StudentEntity> studentEntityList = new LinkedList<>();
List<EnrollStudentEntity> enrollStudentEntityList = new LinkedList<>();
List<AllusersEntity> allusersEntityList = new LinkedList<>();

for (FreshStudentAndStudentModel freshStudentAndStudentModel : list) {

    EnrollStudentEntity enrollStudentEntity = new EnrollStudentEntity();
    StudentEntity studentEntity = new StudentEntity();
    BeanUtils.copyProperties(freshStudentAndStudentModel, studentEntity);
    BeanUtils.copyProperties(freshStudentAndStudentModel, enrollStudentEntity);
    String operator = TenancyContext.UserID.get();
    String studentId = BaseUuidUtils.base58Uuid();
    enrollStudentEntity.setId(BaseUuidUtils.base58Uuid());
    enrollStudentEntity.setStudentId(studentId);
    enrollStudentEntity.setIdentityCardId(freshStudentAndStudentModel.getIdCard());
    enrollStudentEntity.setOperator(operator);
    studentEntity.setId(studentId);
    studentEntity.setIdentityCardId(freshStudentAndStudentModel.getIdCard());
    studentEntity.setOperator(operator);
    studentEntityList.add(studentEntity);
    enrollStudentEntityList.add(enrollStudentEntity);

    AllusersEntity allusersEntity = new AllusersEntity();
    allusersEntity.setId(enrollStudentEntity.getId());
    allusersEntity.setUserCode(enrollStudentEntity.getNemtCode());
    allusersEntity.setUserName(enrollStudentEntity.getName());
    allusersEntity.setSchoolNo(schoolNo);
    allusersEntity.setTelNum(enrollStudentEntity.getTelNum());
    allusersEntity.setPassword(enrollStudentEntity.getNemtCode());  //密码设置为考生号
    allusersEntityList.add(allusersEntity);
}

int c = 100;
int b = enrollStudentEntityList.size() / c;
int d = enrollStudentEntityList.size() % c;

int enResult = 0;
int stuResult = 0;
boolean allResult = false;

for (int e = c; e <= c * b; e = e + c) {
    enResult = enrollStudentDao.insertAll(enrollStudentEntityList.subList(e - c, e));
    stuResult = studentDao.insertAll(studentEntityList.subList(e - c, e));
    allResult = allusersFacade.insertUserList(allusersEntityList.subList(e - c, e));
}
if (d != 0) {
    enResult = enrollStudentDao.insertAll(enrollStudentEntityList.subList(c * b, enrollStudentEntityList.size()));
    stuResult = studentDao.insertAll(studentEntityList.subList(c * b, studentEntityList.size()));
    allResult = allusersFacade.insertUserList(allusersEntityList.subList(c * b, allusersEntityList.size()));
}

if (enResult > 0 && stuResult > 0 && allResult) {
    return 10;
}
return -10;
    }

代码说明:

这样操作,可以避免上面的错误,但是分多次插入,无形中就增加了操作实践,很容易超时。所以这种方法还是不值得提倡的。

再次改进,使用多线程分批导入。在此推荐一下小程序Java精选面试题,内涵大量初级、中级、高级甚至架构师面试题,支持在线刷题。

四、多线程分批导入Mysql

依然使用mybatis的批量导入,不同的是,根据线程数目进行分组,然后再建立多线程池,进行导入。

  @Transactional(rollbackFor = Exception.class)
    public int addFreshStudentsNew(List<FreshStudentAndStudentModel> list, String schoolNo) {
if (list == null || list.isEmpty()) {
    return 0;
}
List<StudentEntity> studentEntityList = new LinkedList<>();
List<EnrollStudentEntity> enrollStudentEntityList = new LinkedList<>();
List<AllusersEntity> allusersEntityList = new LinkedList<>();

list.forEach(freshStudentAndStudentModel -> {
    EnrollStudentEntity enrollStudentEntity = new EnrollStudentEntity();
    StudentEntity studentEntity = new StudentEntity();
    BeanUtils.copyProperties(freshStudentAndStudentModel, studentEntity);
    BeanUtils.copyProperties(freshStudentAndStudentModel, enrollStudentEntity);
    String operator = TenancyContext.UserID.get();
    String studentId = BaseUuidUtils.base58Uuid();
    enrollStudentEntity.setId(BaseUuidUtils.base58Uuid());
    enrollStudentEntity.setStudentId(studentId);
    enrollStudentEntity.setIdentityCardId(freshStudentAndStudentModel.getIdCard());
    enrollStudentEntity.setOperator(operator);
    studentEntity.setId(studentId);
    studentEntity.setIdentityCardId(freshStudentAndStudentModel.getIdCard());
    studentEntity.setOperator(operator);
    studentEntityList.add(studentEntity);
    enrollStudentEntityList.add(enrollStudentEntity);

    AllusersEntity allusersEntity = new AllusersEntity();
    allusersEntity.setId(enrollStudentEntity.getId());
    allusersEntity.setUserCode(enrollStudentEntity.getNemtCode());
    allusersEntity.setUserName(enrollStudentEntity.getName());
    allusersEntity.setSchoolNo(schoolNo);
    allusersEntity.setTelNum(enrollStudentEntity.getTelNum());
    allusersEntity.setPassword(enrollStudentEntity.getNemtCode());  //密码设置为考生号
    allusersEntityList.add(allusersEntity);
});


int nThreads = 50;

int size = enrollStudentEntityList.size();
ExecutorService executorService = Executors.newFixedThreadPool(nThreads);
List<Future<Integer>> futures = new ArrayList<Future<Integer>>(nThreads);

for (int i = 0; i < nThreads; i++) {
    final List<EnrollStudentEntity> EnrollStudentEntityImputList = enrollStudentEntityList.subList(size / nThreads * i, size / nThreads * (i + 1));
    final List<StudentEntity> studentEntityImportList = studentEntityList.subList(size / nThreads * i, size / nThreads * (i + 1));
    final List<AllusersEntity> allusersEntityImportList = allusersEntityList.subList(size / nThreads * i, size / nThreads * (i + 1));

   Callable<Integer> task1 = () -> {
  studentSave.saveStudent(EnrollStudentEntityImputList,studentEntityImportList,allusersEntityImportList);
       return 1;
    };
  futures.add(executorService.submit(task1));
}
executorService.shutdown();
if (!futures.isEmpty() && futures != null) {
    return 10;
}
return -10;
    }

代码说明:

上面是通过应用ExecutorService 建立了固定的线程数,然后根据线程数目进行分组,批量依次导入。一方面可以缓解数据库的压力,另一个面线程数目多了,一定程度会提高程序运行的时间。缺点就是要看服务器的配置,如果配置好的话就可以开多点线程,配置差的话就开小点。

五、小结

通过使用这个操作真是不断的提高了,项目使用技巧也是不错。加油~~ 多线程哦~~

~阅读全文-人机检测~

微信公众号“Java精选”(w_z90110),专注Java技术干货分享!让你从此路人变大神!回复关键词领取资料:如Mysql、Hadoop、Dubbo、Spring Boot等,免费领取视频教程、资料文档和项目源码。微信搜索小程序“Java精选面试题”,内涵3000+道Java面试题!

涵盖:互联网那些事、算法与数据结构、SpringMVC、Spring boot、Spring Cloud、ElasticSearch、Linux、Mysql、Oracle等

评论

分享:

支付宝

微信