Hibernate ——二级缓存

一、Hibernate 二级缓存

1.Hibernate 二级缓存是 SessionFactory 级别的缓存。

2.二级缓存分为两类:

(1)Hibernate内置二级缓存

(2)外置缓存,可配置的,可插拨的,外置缓存中的数据是数据库数据的复制。

3.二级缓存的并发访问策略

(1)两个并发的事务同时访问持久层的缓存的相同数据时,也有可能出现并发问题。

(2)二级缓存可以设定以下 4 中并发访问策略,每一种对应一种事务隔离级别。

  • 非严格读写(Nonstrict-read-write):不保证缓存与数据库中数据的一致性。对应 Read UnCommited 事务隔离级别。
  • 读写型(Read-write):提供 Read Commited 级别的事务隔离级别。
  • 事务型(Transactional):对应 Repeatable Read 级别的事务隔离级别。
  • 只读型(Read-Only):提供 Serializable 级别的数据隔离级别。

4.Hibernate 二级缓存是进程或集群范围内的缓存。是可配置的的插件。这里以 Ehcache 为例。不支持事务型的并发访问策略。

5.配置 Ehcache

(1)添加 Jar 包

Hibernate ——二级缓存

(2)在 Hibernate 配置文件中启用二级缓存并指定适配的缓存适配器。

在 <session-factory> 元素内添加:

<property name="cache.use_second_level_cache">true</property>
<property name="hibernate.cache.region.factory_class">org.hibernate.cache.ehcache.EhCacheRegionFactory</property>

(3)在 Hibernate 配置文件中配置需要使用二级缓存的持久化类,并设置它的二级缓存的并发访问策略。如:

<class-cache class="com.solverpeng.hql.Department" usage="read-write"/>

需要注意的是:<session-factory> 元素内子元素的顺序:property*, mapping*, (class-cache|collection-cache)*, event*, listener*,class-cache 元素必须位于 mapping 节点后。

(4)或者也可以在 hbm 文件中配置,如:

<hibernate-mapping>

    <class name="com.solverpeng.hql.Employee" table="employee" schema="hibernate">
<cache usage="read-write"/>
<id name="empId" column="emp_id"/>
<property name="empName" column="emp_name"/>
<property name="salary" column="salary"/>
<many-to-one name="dept" class="com.solverpeng.hql.Department">
<column name="dept_id_fk" not-null="true"/>
</many-to-one>
</class>
</hibernate-mapping>

6. 测试 Hibernate 二级缓存

(1)查询单个对象缓存

@Test
public void testSecondCache() {
Employee o = (Employee) session.get(Employee.class, 6);
System.out.println(o); transaction.commit();
session.close();
session = factory.openSession();
transaction = session.beginTransaction(); Employee o1 = (Employee) session.get(Employee.class, 6);
System.out.println(o1);
}

查询同一个对象两次,中间关闭 Session 再开启。

没有配置二级缓存:

Hibernate:
select
employee0_.emp_id as emp1_1_0_,
employee0_.emp_name as emp2_1_0_,
employee0_.salary as salary3_1_0_,
employee0_.dept_id_fk as dept4_1_0_
from
hibernate.employee employee0_
where
employee0_.emp_id=?
com.solverpeng.hql.Employee@40dd58f3
Hibernate:
select
employee0_.emp_id as emp1_1_0_,
employee0_.emp_name as emp2_1_0_,
employee0_.salary as salary3_1_0_,
employee0_.dept_id_fk as dept4_1_0_
from
hibernate.employee employee0_
where
employee0_.emp_id=?
com.solverpeng.hql.Employee@40dd58f3

结果:

发送了2条 SQL 语句。

配置二级缓存后(可以在Hibernate 配置文件中配置,也可以在 Employee hbm 配置文件中配置):

Hibernate:
select
employee0_.emp_id as emp1_1_0_,
employee0_.emp_name as emp2_1_0_,
employee0_.salary as salary3_1_0_,
employee0_.dept_id_fk as dept4_1_0_
from
hibernate.employee employee0_
where
employee0_.emp_id=?
com.solverpeng.hql.Employee@40dd58f3
com.solverpeng.hql.Employee@40dd58f3

结果:

只发送了1条 SQL 语句。

(2)集合缓存

@Test
public void testCollectionSecondLevelCache() {
Department department = (Department) session.get(Department.class, 6);
System.out.println(department.getDeptName());
System.out.println(department.getEmps().size()); transaction.commit();
session.close();
session = factory.openSession();
transaction = session.beginTransaction(); Department department2 = (Department) session.get(Department.class, 6);
System.out.println(department2.getDeptName());
System.out.println(department2.getEmps().size()); }

在没有配置二级缓存的情况下:

Hibernate:
select
department0_.dept_id as dept1_0_0_,
department0_.dept_name as dept2_0_0_
from
hibernate.department department0_
where
department0_.dept_id=?
dept-aa
Hibernate:
select
emps0_.dept_id_fk as dept4_0_1_,
emps0_.emp_id as emp1_1_1_,
emps0_.emp_id as emp1_1_0_,
emps0_.emp_name as emp2_1_0_,
emps0_.salary as salary3_1_0_,
emps0_.dept_id_fk as dept4_1_0_
from
hibernate.employee emps0_
where
emps0_.dept_id_fk=?
3
Hibernate:
select
department0_.dept_id as dept1_0_0_,
department0_.dept_name as dept2_0_0_
from
hibernate.department department0_
where
department0_.dept_id=?
dept-aa
Hibernate:
select
emps0_.dept_id_fk as dept4_0_1_,
emps0_.emp_id as emp1_1_1_,
emps0_.emp_id as emp1_1_0_,
emps0_.emp_name as emp2_1_0_,
emps0_.salary as salary3_1_0_,
emps0_.dept_id_fk as dept4_1_0_
from
hibernate.employee emps0_
where
emps0_.dept_id_fk=?
3

结果:

查询了两次 Department,两次Employee

只配置 Department 二级缓存:

<class-cache class="com.solverpeng.hql.Department" usage="read-write"/>
Hibernate:
select
department0_.dept_id as dept1_0_0_,
department0_.dept_name as dept2_0_0_
from
hibernate.department department0_
where
department0_.dept_id=?
dept-aa
Hibernate:
select
emps0_.dept_id_fk as dept4_0_1_,
emps0_.emp_id as emp1_1_1_,
emps0_.emp_id as emp1_1_0_,
emps0_.emp_name as emp2_1_0_,
emps0_.salary as salary3_1_0_,
emps0_.dept_id_fk as dept4_1_0_
from
hibernate.employee emps0_
where
emps0_.dept_id_fk=?
3
dept-aa
Hibernate:
select
emps0_.dept_id_fk as dept4_0_1_,
emps0_.emp_id as emp1_1_1_,
emps0_.emp_id as emp1_1_0_,
emps0_.emp_name as emp2_1_0_,
emps0_.salary as salary3_1_0_,
emps0_.dept_id_fk as dept4_1_0_
from
hibernate.employee emps0_
where
emps0_.dept_id_fk=?
3

结果:

只查询了一次 Department,2次Employee。

说明:

开启 Department 二级缓存后,会对 Department 进行缓存,而与其关联的 emps 不会进行缓存。

配置 Department 二级缓存,同时配置关联的 emps 缓存。

<class-cache class="com.solverpeng.hql.Department" usage="read-write"/>
<collection-cache collection="com.solverpeng.hql.Department.emps" usage="read-write"/>
Hibernate:
select
department0_.dept_id as dept1_0_0_,
department0_.dept_name as dept2_0_0_
from
hibernate.department department0_
where
department0_.dept_id=?
dept-aa
Hibernate:
select
emps0_.dept_id_fk as dept4_0_1_,
emps0_.emp_id as emp1_1_1_,
emps0_.emp_id as emp1_1_0_,
emps0_.emp_name as emp2_1_0_,
emps0_.salary as salary3_1_0_,
emps0_.dept_id_fk as dept4_1_0_
from
hibernate.employee emps0_
where
emps0_.dept_id_fk=?
3
-----------------------------
dept-aa
Hibernate:
select
employee0_.emp_id as emp1_1_0_,
employee0_.emp_name as emp2_1_0_,
employee0_.salary as salary3_1_0_,
employee0_.dept_id_fk as dept4_1_0_
from
hibernate.employee employee0_
where
employee0_.emp_id=?
Hibernate:
select
employee0_.emp_id as emp1_1_0_,
employee0_.emp_name as emp2_1_0_,
employee0_.salary as salary3_1_0_,
employee0_.dept_id_fk as dept4_1_0_
from
hibernate.employee employee0_
where
employee0_.emp_id=?
Hibernate:
select
employee0_.emp_id as emp1_1_0_,
employee0_.emp_name as emp2_1_0_,
employee0_.salary as salary3_1_0_,
employee0_.dept_id_fk as dept4_1_0_
from
hibernate.employee employee0_
where
employee0_.emp_id=?
3

结果:

发送了更多的查询 Employee 的SQL。

说明:

开启集合的二级缓存后,此时会缓存集合中对象的 id ,而不会对集合中的对象进行缓存。若想缓存,需要关联的集合中的对象也开启二级缓存。如:

<class-cache class="com.solverpeng.hql.Department" usage="read-write"/>
<collection-cache collection="com.solverpeng.hql.Department.emps" usage="read-write"/>
<class-cache class="com.solverpeng.hql.Employee" usage="read-write"/>
Hibernate:
select
department0_.dept_id as dept1_0_0_,
department0_.dept_name as dept2_0_0_
from
hibernate.department department0_
where
department0_.dept_id=?
dept-aa
Hibernate:
select
emps0_.dept_id_fk as dept4_0_1_,
emps0_.emp_id as emp1_1_1_,
emps0_.emp_id as emp1_1_0_,
emps0_.emp_name as emp2_1_0_,
emps0_.salary as salary3_1_0_,
emps0_.dept_id_fk as dept4_1_0_
from
hibernate.employee emps0_
where
emps0_.dept_id_fk=?
3
-----------------------------
dept-aa
3

结果:

除 Department 外,关联的 Employee 也被缓存了。

(3)查询缓存(针对 HQL、QBC)

在二级缓存开启的情况下,HQL、QBC 也不能对查询进行缓存。

@Test
public void testQueryCache() {
Query query = session.createQuery("from Employee "); List<Employee> list = query.list();
System.out.println(list.size()); List<Employee> emps2 = query.list();
System.out.println(emps2.size());
}
Hibernate:
select
employee0_.emp_id as emp1_1_,
employee0_.emp_name as emp2_1_,
employee0_.salary as salary3_1_,
employee0_.dept_id_fk as dept4_1_
from
hibernate.employee employee0_
12
Hibernate:
select
employee0_.emp_id as emp1_1_,
employee0_.emp_name as emp2_1_,
employee0_.salary as salary3_1_,
employee0_.dept_id_fk as dept4_1_
from
hibernate.employee employee0_
12

开启查询缓存:

  • 在 Hibernate 配置文件中开启查询缓存
  • 若想启用查询缓存的查询语句,需要调用 Query 或 Criteria 的 setCacheable() 方法。
  • 查询缓存依赖于二级缓存
<property name="hibernate.cache.use_query_cache">true</property>
@Test
public void testQueryCache() {
Query query = session.createQuery("from Employee ");
query.setCacheable(true);
List<Employee> list = query.list();
System.out.println(list.size()); List<Employee> emps2 = query.list();
System.out.println(emps2.size());
}
Hibernate:
select
employee0_.emp_id as emp1_1_,
employee0_.emp_name as emp2_1_,
employee0_.salary as salary3_1_,
employee0_.dept_id_fk as dept4_1_
from
hibernate.employee employee0_
12
12

(4)时间戳缓存区域:时间戳缓存区存放了对于查询结果相关的表进行插入、更新或删除操作的时间戳。Hibernate通过时间戳缓存区来判定被缓存的查询结果是否过期。

@Test
public void testTimStampCache() {
Query query = session.createQuery("from Employee ");
query.setCacheable(true);
List<Employee> list = query.list();
System.out.println(list.size()); Employee employee = (Employee) session.get(Employee.class, 6);
employee.setEmpName("emp@@"); List<Employee> emps2 = query.list();
System.out.println(emps2.size());
}
Hibernate:
select
employee0_.emp_id as emp1_1_,
employee0_.emp_name as emp2_1_,
employee0_.salary as salary3_1_,
employee0_.dept_id_fk as dept4_1_
from
hibernate.employee employee0_
12
Hibernate:
update
hibernate.employee
set
emp_name=?,
salary=?,
dept_id_fk=?
where
emp_id=?
Hibernate:
select
employee0_.emp_id as emp1_1_,
employee0_.emp_name as emp2_1_,
employee0_.salary as salary3_1_,
employee0_.dept_id_fk as dept4_1_
from
hibernate.employee employee0_
12

二、ehcache.xml

<ehcache>
<diskStore path="java.io.tmpdir"/> <defaultCache
maxElementsInMemory="10000"
eternal="false"
timeToIdleSeconds="120"
timeToLiveSeconds="120"
overflowToDisk="true"
/> <cache name="sampleCache1"
maxElementsInMemory="10000"
eternal="false"
timeToIdleSeconds="300"
timeToLiveSeconds="600"
overflowToDisk="true"
/> <cache name="sampleCache2"
maxElementsInMemory="1000"
eternal="true"
timeToIdleSeconds="0"
timeToLiveSeconds="0"
overflowToDisk="false"
/> </ehcache>

1.<diskStore>: 指定一个目录:当 EHCache 把数据写到硬盘上时, 将把数据写到这个目录下.

2.<defaultCache>: 设置缓存的默认数据过期策略

3.<cache> 设定具体的命名缓存的数据过期策略。每个命名缓存代表一个缓存区域

4.缓存区域(region):一个具有名称的缓存块,可以给每一个缓存块设置不同的缓存策略。如果没有设置任何的缓存区域,则所有被缓存的对象,都将使用默认的缓存策略。即:<defaultCache.../>

5.Hibernate在不同的缓存区域保存不同的类/集合。

  • 对于类而言,区域的名称是类名。如:com.atguigu.domain.Customer
  • 对于集合而言,区域的名称是类名加属性名。如com.atguigu.domain.Customer.orders

6.cache 元素的属性

(1)name:设置缓存的名字,它的取值为类的全限定名或类的集合的名字

(2)maxInMemory:设置基于内存的缓存中可存放的对象最大数目

(3)eternal:设置对象是否为永久的,true表示永不过期,此时将忽略timeToIdleSeconds 和 timeToLiveSeconds属性; 默认值是false。

(4)timeToIdleSeconds:设置对象空闲最长时间,以秒为单位, 超过这个时间,对象过期。当对象过期时,EHCache会把它从缓存中清除。如果此值为0,表示对象可以无限期地处于空闲状态。

(5)timeToLiveSeconds:设置对象生存最长时间,超过这个时间,对象过期。如果此值为0,表示对象可以无限期地存在于缓存中. 该属性值必须大于或等于 timeToIdleSeconds 属性值。

(6)overflowToDisk:设置基于内存的缓存中的对象数目达到上限后,是否把溢出的对象写到基于硬盘的缓存中。

三、管理 Session

–Session 对象的生命周期与本地线程绑定

–Session 对象的生命周期与 JTA 事务绑定

–Hibernate 委托程序管理 Session 对象的生命周期

四、批量处理数据

建议通过 JDBC 的方式来进行批量操作。

@Test
public void testBathch() {
session.doWork(new Work() {
@Override
public void execute(Connection connection) throws SQLException {
// 执行批量操作
}
});
}

五、总结

配置 Hibernate 二级缓存的步骤:

1.配置 Hibernate 配置文件

(1)配置启用二级缓存

<property name="cache.use_second_level_cache">true</property>

(2)配置二级缓存使用的产品

<property name="hibernate.cache.region.factory_class">org.hibernate.cache.ehcache.EhCacheRegionFactory</property>

(3)配置对哪些类使用 hibernate 的二级缓存以及并发策略

<class-cache class="com.solverpeng.hql.Department" usage="read-write"/>
<collection-cache collection="com.solverpeng.hql.Department.emps" usage="read-write"/>
<class-cache class="com.solverpeng.hql.Employee" usage="read-write"/>

(4)在 hbm 文件中配置缓存

<set name="emps" inverse="true">
<cache usage="read-write"/>
<key>
<column name="dept_id_fk" not-null="true"/>
</key>
<one-to-many not-found="ignore" class="com.solverpeng.hql.Employee"/>
</set>

2. 对于集合缓存来说,还需要配置集合中的元素对应的持久化类也使用二级缓存! 否则将会多出 n 条 SQL 语句.

3. 查询缓存

(1)在 Hibernate 配置文件中开启查询缓存支持:<property name="cache.use_query_cache">true</property>

(2)调用 Query 或 Criteria 的 setCacheable(true)方法

(3)查询缓存依赖于二级缓存

上一篇:深入理解Java内存模型之系列篇[转]


下一篇:深入理解java内存模型