标签:
基础系统的学生信息维护页面,查询显示学生信息,但每次查询过程都很慢,所以对该部分进行优化.
Hibernate调优是使用SSH框架很常见的问题,一般可以从以下几方面考虑:
一、数据库设计调整
二、主配置参数(ID生成策略,二级缓存,lazy加载)
三、一级缓存管理
四、事务控制策略
五、HQL优化
我们这里的优化,是对hql语句的优化,考虑使用select newmap代替hibernate的级联查询。
先说说实体关联关系是怎么样的:
在学生实体Student中,classes是作为一个关联实体而存在的.
在数据库中表现为tb_student表中有一个classesId字段.
在班级实体classes中,institution(专业)也是作为一个关联实体而存在的.
在数据库表中表现为tb_classes表中有一个institutionId字段.
同样地,institution实体中,nodeLevel(学校层次)也是作为一个关联实体而存在的.
在数据库表中表现为tb_institution表中有一个levelId字段.
介绍完实体级联关系和数据库表结构,我们来看看真实数据.
学生表中有1.8万条数据,如下所示.
我做的是分页查询,每页10条学生信息.
优化前:
查询语句:
<span style="white-space:pre"> </span>String hql ="From Student where isDelete =:isDelete ";
查询过程:
发送150条select语句,用时28秒.
11:43:37,290 INFO [stdout] (http-localhost/127.0.0.1:8080-1)28690毫秒
优化后:
查询语句:
String hql="select new map (S.id as id," + "S.name as name," + "S.comment as comment," + "S.remark as remark," + "S.accountAddress as accountAddress," + "S.code as code," + "S.email as email," + "S.identityCardID as identityCardID," + "S.classes.className as className," + "S.classes.institution.institutionName as institutionName," + "S.classes.institution.parentInstitution.institutionName as parentInstitution," + "S.classes.institution.schoolLevel.levelName as levelName," + "S.classes.institution.id as institutionId," + "P.professionalName as professionalName " + ") from Student as S left join S.professional as P " +" where S.isDelete=:isDelete ";
查询过程:
共发送1条语句,用时38毫秒.
17:45:38,054 INFO [stdout] (http-localhost/127.0.0.1:8080-6)Hibernate: select count(*) as col_0_0_ from TB_Student student0_ left outerjoin TB_Professional profession1_ on student0_.professionalId=profession1_.id wherestudent0_.isDelete=?
17:45:38,080 INFO [stdout] (http-localhost/127.0.0.1:8080-6)Hibernate: select student0_.id as col_0_0_, student0_.name as col_1_0_,student0_.comment as col_2_0_, student0_.remark as col_3_0_,student0_.accountAddress as col_4_0_, student0_.code as col_5_0_,student0_.email as col_6_0_, student0_.entranceDate as col_7_0_,student0_.graduateSchool as col_8_0_, student0_.identityCardID as col_9_0_,classes2_.className as col_10_0_, institutio4_.InstitutionName as col_11_0_,institutio7_.InstitutionName as col_12_0_, schoolleve10_.levelName ascol_13_0_, classes2_.institutionId as col_14_0_, profession1_.professionalNameas col_15_0_ from TB_Student student0_ left outer join TB_Professionalprofession1_ on student0_.professionalId=profession1_.id, TB_Classes classes2_,TB_Institution institutio4_, TB_Institution institutio7_, TB_SchoolLevelschoolleve10_ where student0_.calssesId=classes2_.id andclasses2_.institutionId=institutio4_.id and institutio4_.pId=institutio7_.idand institutio4_.levelId=schoolleve10_.id and student0_.isDelete=? limit ?
17:45:38,091 INFO [stdout] (http-localhost/127.0.0.1:8080-6)38毫秒
关于优化的分析请参考后续博客.
版权声明:本文为博主原创文章,未经博主允许不得转载。
标签:
原文地址:http://blog.csdn.net/zhuanzhe117/article/details/47045507