Mybatis缓存解析

最新推荐文章于 2022-02-23 00:58:19 发布
最新推荐文章于 2022-02-23 00:58:19 发布
阅读量285 收藏
点赞数
分类专栏： Mybatis 文章标签： mybatis
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/D578332749/article/details/81368337
版权
一、Mybatis的缓存概述

声明：本文纯属个人兴趣探究，其中一些概念、理论以及结论并不一定正确，如果有错误，请包涵并告知。

Mybatis作为持久层框架，要提升性能必然离不开缓存，Mybatis为我们提供了两级缓存，称为一级缓存和二级缓存，另外Mybatis还支持自定义缓存，本文仅探究一级缓存和二级缓存。

一级缓存和二级缓存是分别作用于不同的工作域，要了解缓存就需要明白Mybatis工作时基本的实例。

SqlSession：从字面可以看出，这个对象负责一次与数据库的会话，会话结束后即会被销毁，然后数据会进行提交或回滚，可以理解为它代表着一次事务。

SqlSessionFactory：顾名思义，就是生产SqlSession的工厂，所有的SqlSession都从这里获得，一个数据库连接对应着一个SqlSessionFactory，所以如果系统中需要同时连接多个数据库，那么就需要保存多个SqlSessionFactory的实例。

Mapper：Mapper对应的是一个SQL映射文件，包含增删改查等具体操作，Mybatis的二级缓存便是对应一个个的Mapper。

再来看Mybatis的缓存。

一级缓存

一级缓存是SqlSession级别的，作用在一个SqlSession内，即在一个SqlSession内，重复提交完全一样的sql，不会再重复执行，而是从缓存中去获取，SqlSession关闭后对应于这个SqlSession的缓存就会被清理，一级缓存默认开启。

我们通过一些简单的测试代码来了解Mybatis的一级缓存。

1、执行两次相同的查询，观察SQL执行情况

测试代码：

public Map<String, Object> queryCustomer(Integer customerId) throws IOException{
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
        TestMapper testMapper = sqlSession.getMapper(TestMapper.class);
        //第一次查询
        System.out.println("-----------queryCustomer1-----------");
        Map<String, Object> customer1 = testMapper.queryCustomer(customerId);
        System.out.println("-----------queryCustomer2-----------");
        //第二次查询，执行完全相同查询语句，来测试是否会从缓存中获取数据
        Map<String, Object> customer2 = testMapper.queryCustomer(customerId);
        return customer2;
    } finally {
        sqlSession.close();
    }
}

输出结果：

T E S T S
-------------------------------------------------------
Running com.study.mybatis.service.TestServiceTester
Logging initialized using 'class org.apache.ibatis.logging.stdout.StdOutImpl' adapter.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
-----------queryCustomer1-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Opening JDBC Connection
Created connection 188576144.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
-----------queryCustomer2-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Returned connection 188576144 to pool.
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.658 sec

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

结论：可以看到，再次执行相同查询语句时，Mybatis就从一级缓存中去获取，不会重复执行SQL。

2、在两次完全相同的查询中插入一个修改的操作，查询与修改是同一张数据表。

测试代码：

public Map<String, Object> queryCustomer(Integer customerId) throws IOException{
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
        TestMapper testMapper = sqlSession.getMapper(TestMapper.class);
        //第一次查询
        System.out.println("-----------queryCustomer1-----------");
        Map<String, Object> customer1 = testMapper.queryCustomer(customerId);
        System.out.println("-----------updateCustomer-----------");
        Map<String, Object> map = new HashMap<>();
        map.put("customerId", customerId);
        map.put("birthday", "1992-09-01");
        //进行一次update操作，select和update是同一张数据表
        testMapper.updateCustomerBirthday(map);
        System.out.println("-----------queryCustomer2-----------");
        //第二次查询
        Map<String, Object> customer2 = testMapper.queryCustomer(customerId);
        sqlSession.commit();
    } catch (Exception e) {
        sqlSession.rollback();
    } finally {
        sqlSession.close();
    }
    return null;
}

输出结果：

T E S T S
-------------------------------------------------------
Running com.study.mybatis.service.TestServiceTester
Logging initialized using 'class org.apache.ibatis.logging.stdout.StdOutImpl' adapter.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
-----------queryCustomer1-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Opening JDBC Connection
Created connection 188576144.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
-----------updateCustomer-----------
==>  Preparing: update tb_customer_info t set t.customerbirthday = ? where t.customerid = ? 
==> Parameters: 1992-09-01(String), 269(Integer)
<==    Updates: 1
-----------queryCustomer2-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1992-09-01, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
Rolling back JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Returned connection 188576144 to pool.
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.669 sec

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

结论：在两次完全相同的查询操作中插入一次update操作，update和select操作均为同一张数据表，在update操作后Mybatis就会将该SqlSession的一级缓存清理，所以在第二次执行相同的SQL时没有在一级缓存中找到结果，再次执行了SQL。

3、在两次完全相同的查询中插入一个修改的操作，查询与修改不是同一张数据表。

测试代码：

public Map<String, Object> queryCustomer(Integer customerId) throws IOException{
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
        TestMapper testMapper = sqlSession.getMapper(TestMapper.class);
        System.out.println("-----------queryCustomer1-----------");
        //第一次查询
        Map<String, Object> customer1 = testMapper.queryCustomer(customerId);
        System.out.println("-----------updateCustomer-----------");
        Map<String, Object> map = new HashMap<>();
        map.put("zoneCode", 1);
        map.put("zoneName", "区域1");
        //进行一次update操作，select和update不是同一张数据表
        testMapper.updateZoneName(map);
        System.out.println("-----------queryCustomer2-----------");
        //第二次查询
        Map<String, Object> customer2 = testMapper.queryCustomer(customerId);
        sqlSession.commit();
    } catch (Exception e) {
        sqlSession.rollback();
    } finally {
        sqlSession.close();
    }
    return null;
}

输出结果：

 T E S T S
-------------------------------------------------------
Running com.study.mybatis.service.TestServiceTester
Logging initialized using 'class org.apache.ibatis.logging.stdout.StdOutImpl' adapter.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
-----------queryCustomer1-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Opening JDBC Connection
Created connection 188576144.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
-----------updateZone-----------
==>  Preparing: update tb_initzone t set t.zone_name = ? where t.zone_code = ? 
==> Parameters: 区域1(String), 1(Integer)
<==    Updates: 1
-----------queryCustomer2-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
Committing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Returned connection 188576144 to pool.
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.643 sec

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

结论：在两次完全相同的查询操作中插入一次update操作，update和select操作不为同一张数据表，在update操作后Mybatis仍然会将该SqlSession的一级缓存清理，也就是在SqlSession中只要涉及到数据改动，Mybatis就会将其一级缓存清理。

二级缓存

二级缓存是Mapper级别的，每一个Mapper（对应于一份SQL映射文件）都有自己的一份二级缓存，并且是所有SqlSession共享的，不会随着数据库会话的关闭而被清理，启用二级缓存后，在一定的时间内，全局的SqlSession使用该Mapper执行相同的查询操作会优先从该Mapper的二级缓存中查找获取，该Mapper执行增删改操作后会将其二级缓存清空。二级缓存默认关闭，需要手动在配置文件中开启，Mybatis为二级缓存提供了一些配置项，可以针对实际需求自行配置。

开启二级缓存

1、Mybatis配置文件settings节点下设置cacheEnabled，开启二级缓存

<setting name="cacheEnabled" value="true" />

2、SQL映射文件中增加cache节点，声明此映射文件中配置的sql需要使用二级缓存

<mapper namespace="com.study.mybatis.mapper.TestMapper">
    <cache/>
</mapper>

通过以上两步即可开启使用Mybatis的二级缓存。

我们通过一些简单的程序来测试Mybatis的二级缓存。

1、用两个SqlSession先后执行相同的查询语句，观察是否会从缓存中获取数据

业务层代码：

public Map<String, Object> query1(Integer customerId) throws IOException{
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
        TestMapper testMapper = sqlSession.getMapper(TestMapper.class);
        System.out.println("-----------query1-----------");
        return testMapper.queryCustomer(customerId);
    } finally {
        sqlSession.close();
    }
}

public Map<String, Object> query2(Integer customerId) throws IOException{
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
        TestMapper testMapper = sqlSession.getMapper(TestMapper.class);
        System.out.println("-----------query2-----------");
        return testMapper.queryCustomer(customerId);
    } finally {
        sqlSession.close();
    }
}

测试代码：

@Test
public void testSecondCache() {
    TestService service = new TestService();
    try {
        //第一次查询
        Map<String, Object> list1 = service.query1(269);
        //第二次查询
        Map<String, Object> list2 = service.query2(269);
    } catch (IOException e) {
        e.printStackTrace();
    }
}

输出结果：

 T E S T S
-------------------------------------------------------
Running com.study.mybatis.service.TestServiceTester
Logging initialized using 'class org.apache.ibatis.logging.stdout.StdOutImpl' adapter.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
-----------query1-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Opening JDBC Connection
Created connection 188576144.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Returned connection 188576144 to pool.
-----------query2-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.5
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.919 sec

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

结论：在query1时Mybatis执行了SQL，而query2时没有执行SQL，而是从缓存中获取到了数据，说明二级缓存已经成功开启。

2、在两次查询中增加一次修改操作，观察query2是否会再次执行SQL，修改的表与查询的表不是同一张数据表

业务层代码：

public int update1() throws IOException {
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
        TestMapper testMapper = sqlSession.getMapper(TestMapper.class);
        System.out.println("-----------update-----------");
        Map<String, Object> map = new HashMap<>();
        map.put("zoneCode", 1);
        map.put("zoneName", "区域1");
        int line = testMapper.updateZoneName(map);
        sqlSession.commit();
        return line;
    } catch (Exception e) {
        sqlSession.rollback();
    } finally {
        sqlSession.close();
    }
    return 0;
}
/*query1与query2无变动*/

测试代码：

@Test
public void testSecondCache() {
    TestService service = new TestService();
    try {
        Map<String, Object> list1 = service.query1(269);
        int line = service.update1();
        Map<String, Object> list2 = service.query2(269);
    } catch (IOException e) {
        e.printStackTrace();
    }
}

输出结果：

T E S T S
-------------------------------------------------------
Running com.study.mybatis.service.TestServiceTester
Logging initialized using 'class org.apache.ibatis.logging.stdout.StdOutImpl' adapter.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
-----------query1-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Opening JDBC Connection
Created connection 188576144.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Returned connection 188576144 to pool.
-----------update1-----------
Opening JDBC Connection
Checked out connection 188576144 from pool.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
==>  Preparing: update tb_initzone t set t.zone_name = ? where t.zone_code = ? 
==> Parameters: 区域1(String), 1(Integer)
<==    Updates: 1
Committing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Returned connection 188576144 to pool.
-----------query2-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Opening JDBC Connection
Checked out connection 188576144 from pool.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@b3d7190]
Returned connection 188576144 to pool.
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.685 sec

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

结论：在执行了update操作后，query2再次执行了SQL，说明修改操作触发Mybatis清理二级缓存，即使修改与查询的不是同一张表。

3、在Mapper1中查询，然后在Mapper2中进行修改操作，然后再用Mapper执行相同查询，观察query2时是否会再次执行SQL，修改与查询的不是同一张表

业务层代码：

public int update2() throws IOException {
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
        //注意，此处获取的是Test2Mapper的实例
        Test2Mapper test2Mapper = sqlSession.getMapper(Test2Mapper.class);
        System.out.println("-----------update2-----------");
        Map<String, Object> map = new HashMap<>();
        map.put("zoneCode", 1);
        map.put("zoneName", "区域1");
        int line = test2Mapper.updateZoneName(map);
        sqlSession.commit();
        return line;
    } catch (Exception e) {
        sqlSession.rollback();
    } finally {
        sqlSession.close();
    }
    return 0;
}
/*query1与query2无变动*/

测试代码：

@Test
public void testSecondCache() {
    TestService service = new TestService();
    try {
        Map<String, Object> list1 = service.query1(269);
        int line = service.update2();
        Map<String, Object> list2 = service.query2(269);
    } catch (IOException e) {
        e.printStackTrace();
    }
}

输出结果：

 T E S T S
-------------------------------------------------------
Running com.study.mybatis.service.TestServiceTester
Logging initialized using 'class org.apache.ibatis.logging.stdout.StdOutImpl' adapter.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
-----------query1-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Opening JDBC Connection
Created connection 1873859565.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Returned connection 1873859565 to pool.
-----------update2-----------
Opening JDBC Connection
Checked out connection 1873859565 from pool.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
==>  Preparing: update tb_initzone t set t.zone_name = ? where t.zone_code = ? 
==> Parameters: 区域1(String), 1(Integer)
<==    Updates: 1
Committing JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Returned connection 1873859565 to pool.
-----------query2-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.5
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.653 sec

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

结论：从执行的结果来看，query1后执行update2，再执行query2，query2仍然是从缓存中在获取数据，TestMapper的缓存并没有因为Test2Mapper中的update2而被清理。

4、在Mapper1中查询，然后在Mapper2中进行修改操作，然后再用Mapper执行相同查询，观察query2时是否会再次执行SQL，修改与查询的是同一张表

业务层代码：

public int update3() throws IOException {
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
        //注意，此处获取的是Test2Mapper的实例
        Test2Mapper test2Mapper = sqlSession.getMapper(Test2Mapper.class);
        System.out.println("-----------update3-----------");
        Map<String, Object> map = new HashMap<>();
        map.put("customerid", 269);
        map.put("birthday", "1993-09-08");
        int line = test2Mapper.updateCustomerBirthday(map);
        sqlSession.commit();
        return line;
    } catch (Exception e) {
        sqlSession.rollback();
    } finally {
        sqlSession.close();
    }
    return 0;
}
/*query1与query2无变动*/

测试代码：

@Test
public void testSecondCache() {
    TestService service = new TestService();
    try {
        Map<String, Object> list1 = service.query1(269);
        int line = service.update3();
        Map<String, Object> list2 = service.query2(269);
    } catch (IOException e) {
        e.printStackTrace();
    }
}

输出结果：

 T E S T S
-------------------------------------------------------
Running com.study.mybatis.service.TestServiceTester
Logging initialized using 'class org.apache.ibatis.logging.stdout.StdOutImpl' adapter.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
-----------query1-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.0
Opening JDBC Connection
Created connection 1873859565.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
==>  Preparing: select * from tb_customer_info t where t.customerid = ? 
==> Parameters: 269(Integer)
<==    Columns: CUSTOMERID, CUSTOMERNAME, CUSTOMERCERTTYPE, CUSTOMERCERTNO, CUSTOMERTEL, CUSTOMERMOBILE, CUSTOMERBIRTHDAY, CUSTOMEREMAIL, CUSTOMERADDR, CUSTOMERDEPT, CUSTOMERSEX, CUSTOMERNATION, CUSTOMERBIRTHPLACE, CUSTOMEREDUCATION, CUSTOMERINDUSTRYNO, CUSTOMER_PTO, CUSTOMER_TYPE, CUSTOMER_STATE, CARDTYPEID, INIT_MONEY, FISBIND, FBANKNO, QINS, HOUSE_ID
<==        Row: 269, 黄锦华, 13_000001, null, null, null, 1900-01-0, null, null, 00010015, 19_000002, null, null, null, 6000252, null, null, 1, 16, 0, 0, null, null, null
<==      Total: 1
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Returned connection 1873859565 to pool.
-----------update3-----------
Opening JDBC Connection
Checked out connection 1873859565 from pool.
Setting autocommit to false on JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
==>  Preparing: update tb_customer_info t set t.customerbirthday = ? where t.customerid = ? 
==> Parameters: 1993-09-08(String), 269(Integer)
<==    Updates: 1
Committing JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Resetting autocommit to true on JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Closing JDBC Connection [oracle.jdbc.driver.T4CConnection@6fb0d3ed]
Returned connection 1873859565 to pool.
-----------query2-----------
Cache Hit Ratio [com.study.mybatis.mapper.TestMapper]: 0.5
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.693 sec

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

结论：从一系列测试结果可以看出，Mybatis的缓存和具体的数据表没有关系，TestMapper和Test2Mapper之间不是共享缓存的，Test2Mapper中进行的修改操作，不会影响到TestMapper的缓存，我们在使用就需要小心，因为这个有可能会导致脏数据的发生，应该避免一个数据表在多个Mapper中被操作。

总结

通过一系列的测试可以看出，Mybatis的二级缓存是对应于Mapper的，每个Mapper都有自己的二级缓存，在同一个Mapper下进行增删改操作，会触发清理其二级缓存的动作，如果在一个Mapper中先查询出一份数据，在另一个Mapper中进行增删改操作，而后再用前一个Mapper执行同样的查询动作，这时就出现脏数据，所以我们在使用Mybatis时应该避免在多个Mapper中操作同一个数据表。

在Mybatis初始化时会执行Mapper中的查询语句缓存数据？

之前听到过一种说法，就是在Mybatis初始化时会执行Mapper中的查询语句缓存数据，一直半信半疑，因为大多数查询语句都是带参数的，如何进行预加载呢？

于是通过简单代码进行了测试，我特意在TestMapper.xml中放置了一条无需任何参数的查询语句，TestMapper.java中也有对应的抽象函数，以观察它是否会被预先加载并缓存。

业务代码：

/*仅开启了一个数据库会话，以初始化Mybatis的一系列实例*/
public void load() throws IOException{
    SqlSession sqlSession = Context.getSqlSessionFactory().openSession();
    try {
    } finally {
        sqlSession.close();
    }
}

测试代码：

@Test
public void testLoad() {
    TestService service = new TestService();
    try {
        System.out.println("-----------load-----------");
        service.load();
    } catch (IOException e) {
        e.printStackTrace();
    }
}

输出结果：

 T E S T S
-------------------------------------------------------
Running com.study.mybatis.service.TestServiceTester
-----------load-----------
Logging initialized using 'class org.apache.ibatis.logging.stdout.StdOutImpl' adapter.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
PooledDataSource forcefully closed/removed all connections.
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.484 sec

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

结论：从测试结果来看，Mybatis在初始化SqlSessionFactory时不会自动执行SQL映射文件中的查询语句，开启SqlSession时也不会执行，只有明确调用某个查询动作时才会执行，并进行缓存，也许是需要进行手动设置，暂时不会用到，便不再继续深入了。

imdongrui
关注
  • 0
    点赞
  • 踩
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
专栏目录
Mybaits 中的缓存
weixin_42248417的博客
04-29 217
1。一级缓存，放在sqlSession中，及如果使用同一个sqlSession对象。 图1 同一个缓存 图2不同缓存 当查询想要跳过缓存，则可以用sqlSession.clearCache()来清空缓存。 如果使用sqlSession执行修改操作，此修改操作会把第一次查询的缓存失...
mybatis学习笔记（五）
qq_54898506的博客
03-13 69
mybatis学习笔记（五） 动态SQL 什么是动态SQL：动态SQL就是指根据不同的条件生成不同的SQL语句 利用动态SQL这一特性可以彻底摆脱这种痛苦 if choose (when, otherwise) trim (where, set) foreach 搭建环境 CREATE TABLE `blog`( `id` VARCHAR(50) NOT NULL COMMENT '博客id', `title` VARCHAR(100) NOT NULL COMMENT '博客标题', `author`
mybaits缓存详解
qq_41537781的博客
05-26 146
说明：mybatis提供查询缓存，用于减轻数据库压力，提高数据库性能 mybatis提供一级缓存和二级缓存 一级缓存：sqlsession级别的缓存 在操作数据库时，需要构造sqlsession对象，在对象中有一个数据结构（HashMap）用于存储缓存数据 不同的sqlsession之间的缓存区域是互相不影响的。 二级缓存：mapper级别的缓存 　　　　　多个sqlsession去操作同一个mapper的sql语句，多个s...
MyBatis的缓存分析
dhoe3597的博客
07-14 119
一：MyBatis缓存简介 MyBatis支持声明式数据缓存（declarative data caching）。当一条SQL语句被标记为“可缓存”后，首次执行它时从数据库获取的所有数据会被存储在一段高速缓存中，今后执行这条语句时就会从高速缓存中读取结果，而不是再次命中数据库。MyBatis提供了默认下基于Java HashMap的缓存实现，以及用于与OSCache、Ehcache...
针对mybatis的缓存理解
qq_34921492的博客
07-01 564
今天看了关于mybatis缓存配置的相关文章，主要有以下理解，    1、mybatis分别有一级、二级缓存，即已经设定好的缓存模式，其中，一级缓存指每个sqlsession分别有一个缓存，针对相应情况进行数据的缓存，也是默认开启的功能；而针对二级缓存需要在mybatis的sql映射文件中加入&lt;cache /&gt;语句开启，并且相应的model类要实现Serializable接口，如果只配...
mybatis缓存源码解析
04-23
个人学习mybatis源码总结
Mybatis 缓存原理及失效情况解析
08-25
Mybatis 缓存原理及失效情况解析 Mybatis 缓存是指将经常查询的数据存于内存中，以提高查询效率，解决高并发系统的性能问题。Mybatis 缓存分为一级缓存和二级缓存两种。 一级缓存，也称为本地缓存，指的是...
MyBatis缓存功能原理及实例解析
08-19
MyBatis缓存功能原理及实例解析 MyBatis缓存功能原理及实例解析是MyBatis框架中非常重要的一部分，它可以极大的提升查询效率，解决高并发系统的性能问题。缓存机制可以将用户经常查询的数据放在缓存（内存）中，...
MyBatis缓存实现原理及代码实例解析
09-07
MyBatis提供了两种级别的缓存：一级缓存（本地缓存）和二级缓存（全局缓存）。 一级缓存是SqlSession级别的，它默认始终开启。当在一个SqlSession中执行SQL查询时，查询结果会被存储在本地缓存中。如果后续请求相同...
【MyBatis源码全面解析】MyBatis一二级缓存介绍
08-30
MyBatis缓存 MyBatis中的缓存分为两种：一级缓存和二级缓存。 一级缓存 一级缓存是SqlSession级别的缓存，对于相同的查询，会从缓存中返回结果而不是查询数据库。当openSession()方法运行完毕或者主动调用了...
Mybatis开启日志
weixin_47620760的博客
03-05 4455
日志工厂： 如果一个数据库操作出现了异常，我们需要排错，日志就是最好的助手 曾经： sout、debug 现在： 日志工场 SLF4J LOG4J LOG4J2 JDK_LOGGING COMMONS_LOGGING STDOUT_LOGGING NO_LOGGING 在Mybatis中具体使用哪一个日志实现，在设置中设定 STDOUT_LOGGING 标准日志输出 在mybatis核心配置文件中，配置我们的日志 <settings> <!-- 标准日志工厂实现--&
mybatis详解（2）
weixin_45351103的博客
02-23 360
mybatis详解（2）1 #和$的区别1.1 例子1.1.1 第一步：在接口中写方法1.1.2 第二步：写sql语句1.1.3 第三步：测试1.1.4 补充：如果在第二步用#{price}代替${price}2 缓存问题2.1 用一个查询测试2.1.1 逐行解释2.2 第二次测试2.3 第三次测试2.4 一级缓存2.5 二级缓存2.5.1 设置二级缓存2.5.2 二级缓存的效果2.5.3 二级缓存需要注意的事 1 #和$的区别 # $ PreparedStatement Stateme
MBS12: mybatis的一级缓存与二级缓存
qq_44700366的博客
09-03 231
12.1 简介 ​ 为了降低系统开销，减少与数据库的交互，mybatis会把一些需要经常使用的、不常变更的数据缓存在本地。每次进行数据访问时，仅仅只需要访问本地换粗即可，当然缓存仅仅用于查询，并不适合于增删改，毕竟没有意义。 缓存分为一级缓存与二级缓存 ，一级缓存是 SqlSession级别的 ，与SqlSession的生命周期一致。而 二级缓存则是namespace级别 的，与单个接口级别一致。当然，mybatis提供了 自定义二级缓存 的方法。 12.2 一级缓存 12.2.1 简介 ​ 正如前面介绍
MyBatis学习（七）
weixin_33923762的博客
10-18 109
本课程对应视频教程：http://edu.51cto.com/sd/3ec2c 1、高级查询 MyBatis作为一个ORM框架，也对sql的高级查询做了支持，这里以用户，订单，订单详情，商品为例讲解 案例说明：此案例的业务关系是用户、订单、订单详情、商品之间的关系、其中 一个订单只能属于一个人 一个订单可以有多个订单详情 一个订单详情中包含一个商品信息 他们的关系是： 订单和人是一对一的关系 订...
Setting autocommit to false on JDBC Connection 自动提交失败
qq_42799475的博客
10-25 4968
今天在执行mybstis的测试时，明明已经写好了插入语句但是数据库没有插入数据成功 这段报错的意思是“自动提交失败”，原因是我没有提交事务，后面加上一句 sqlSession.commit()就可以了 也可以在 SqlSession session = MybatisUtil.getSqlSessionFactory().openSession(); 的openSession()方法里加上一个 ...
使用Mybatis添加用户时报：Setting autocommit to false on JDBC Connection
Daneliya的博客
01-24 3022
2019-01-24 15:18:13,391 0 [ main] DEBUG ache.ibatis.logging.LogFactory - Logging initialized using 'class org.apache.ibatis.logging.log4j.Log4jImpl' adapter. 2019-01-24 15:18:13,632 2...
【mybatis】完全解读mybatis JDBC事务
厚积薄发者，轻舟万重山
07-29 7824
前言         mybatis源码版本：3.2.7 一、mybatis的发动机SqlSessionManager
Mybatis 缓存失效的几种情况
A08110123的博客
08-09 333
1　不在同一个sqlSession对象中 下面比较下载同一个sqlSession和不在同一sqlSession下面的两种情况： 同一sqlSession: @Test public final void testQueryClazzById() { SqlSession session = sqlSessionFactory.openSession(...
mybatis mysql 调用存储过程 多个返回值_MyBatis缓存机制（一级缓存，二级缓存）...
weixin_39781326的博客
11-23 136
写在前面：2020年面试必备的Java后端进阶面试题总结了一份复习指南在Github上，内容详细，图文并茂，有需要学习的朋友可以Star一下！GitHub地址：abel-max/Java-Study-Note一，MyBatis一级缓存（本地缓存）My Batis 一级缓存存在于 SqlSession 的生命周期中，是SqlSession级别的缓存。在操作数据库时需要构造SqlSession对象，在...
mybatis源码解析
最新发布
10-22
Mybatis是一个轻量级的Java持久...总之，Mybatis的源码解析涉及多个关键模块的实现，包括配置解析、SQL解析、SQL执行、连接池、缓存、事务管理等。通过了解这些模块的实现原理，我们可以更好地理解和使用Mybatis框架。

“相关推荐”对你有帮助么？

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值