Mybatis一级缓存和结合Spring Framework后失效的源码探究

　　1.在下面的案例中，执行两次查询控制台只会输出一次 SQL 查询：

mybatis-config.xml

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE configuration
        PUBLIC "-//mybatis.org//DTD Config 3.0//EN"
        "http://mybatis.org/dtd/mybatis-3-config.dtd">
<configuration>
    <environments default="development">
        <environment id="development">
            <transactionManager type="JDBC"/>
            <dataSource type="POOLED">
                <property name="driver" value="com.mysql.jdbc.Driver"/>
                <property name="url" value="jdbc:mysql://localhost:3306/xxx?useUnicode=true&amp;characterEncoding=utf-8&amp;autoReconnect=true"/>
                <property name="username" value="xxx"/>
                <property name="password" value="xxx"/>
            </dataSource>
        </environment>
    </environments>
    <mappers>
        <mapper resource="com/hrh/mapper/PersonMapper.xml"/>
    </mappers>
</configuration>

PersonMapper.xml
<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE mapper PUBLIC "-//mybatis.org//DTD Mapper 3.0//EN" "http://mybatis.org/dtd/mybatis-3-mapper.dtd" >
<mapper namespace="com.hrh.mapper.PersonMapper">
    <resultMap id="BaseResultMap" type="com.hrh.bean.Person">
        <id column="id" property="id" jdbcType="BIGINT"/>
        <result column="name" property="name" jdbcType="VARCHAR"/>
        <result column="age" property="age" jdbcType="BIGINT"/>
    </resultMap>
    <sql id="Base_Column_List">
    id, name, age
    </sql>
    <select id="list" resultType="com.hrh.bean.Person">
        select
        <include refid="Base_Column_List"/>
        from tab_person
    </select>
</mapper>

public interface PersonMapper {
     List<Person> list();
}

        String resource = "mybatis-config2.xml";
        InputStream inputStream = Resources.getResourceAsStream(resource);
        SqlSessionFactory sqlSessionFactory =
                new SqlSessionFactoryBuilder().build(inputStream);
        SqlSession sqlSession = sqlSessionFactory.openSession();//开启会话
        PersonMapper mapper = sqlSession.getMapper(PersonMapper.class);
        mapper.list();
        mapper.list();

 

 　　之所以会出现这种情况，是因为 Mybatis 存在一级缓存导致的，下面 debug 探究下内部流程：

　　（1）mapper.list() 会进入 MapperProxy#invoke()：参数 proxy是一个代理对象（每个 Mapper 接口都会被转换成一个代理对象），里面包含会话 sqlSession、接口信息、方法信息；method 是目标方法（当前执行的方法），它里面包含了所属的哪个类（接口）、方法名、返回类型（List、Map、void 或其他）、参数类型等；args 是参数；

  public Object invoke(Object proxy, Method method, Object[] args) throws Throwable {
    try {
      if (Object.class.equals(method.getDeclaringClass())) {
        return method.invoke(this, args);
      } else if (isDefaultMethod(method)) {
        return invokeDefaultMethod(proxy, method, args);
      }
    } catch (Throwable t) {
      throw ExceptionUtil.unwrapThrowable(t);
    }
    //从方法缓存methodCache中获取到方法的信息：比如方法名、类型（select、update等）、返回类型
    //如果获取中没有MapperMethod，则创建一个并放入methodCache中
    final MapperMethod mapperMethod = cachedMapperMethod(method);
    //执行查询SQL并返回结果
    return mapperMethod.execute(sqlSession, args);
  }

　　

　　cacheMapperMethod：MapperMethod 包含方法名、类型（select、update等）、返回类型等信息

  private MapperMethod cachedMapperMethod(Method method) {
    //缓存中获取
    MapperMethod mapperMethod = methodCache.get(method);
    //没有则创建一个对象并放入缓存中供下次方便取用
    if (mapperMethod == null) {
      mapperMethod = new MapperMethod(mapperInterface, method, sqlSession.getConfiguration());
      methodCache.put(method, mapperMethod);
    }
    return mapperMethod;
  }

　　（2）MapperMethod#execute() 根据 SQL 类型进入不同的查询方法

  public Object execute(SqlSession sqlSession, Object[] args) {
    //返回结果
    Object result;
    //判断语句类型
    switch (command.getType()) {
      case INSERT: {//插入语句
      Object param = method.convertArgsToSqlCommandParam(args);
        result = rowCountResult(sqlSession.insert(command.getName(), param));
        break;
      }
      case UPDATE: {//更新语句
        Object param = method.convertArgsToSqlCommandParam(args);
        result = rowCountResult(sqlSession.update(command.getName(), param));
        break;
      }
      case DELETE: {//删除语句
        Object param = method.convertArgsToSqlCommandParam(args);
        result = rowCountResult(sqlSession.delete(command.getName(), param));
        break;
      }
      case SELECT://查询语句
        //返回空的查询
        if (method.returnsVoid() && method.hasResultHandler()) {
          executeWithResultHandler(sqlSession, args);
          result = null;
          //返回List的查询
        } else if (method.returnsMany()) {
          result = executeForMany(sqlSession, args);
          //返回Map的查询
        } else if (method.returnsMap()) {
          result = executeForMap(sqlSession, args);
          //返回游标的查询
        } else if (method.returnsCursor()) {
          result = executeForCursor(sqlSession, args);
        } else {
          Object param = method.convertArgsToSqlCommandParam(args);
          result = sqlSession.selectOne(command.getName(), param);
        }
        break;
      case FLUSH:
        result = sqlSession.flushStatements();
        break;
      default:
        throw new BindingException("Unknown execution method for: " + command.getName());
    }
    if (result == null && method.getReturnType().isPrimitive() && !method.returnsVoid()) {
      throw new BindingException("Mapper method '" + command.getName() 
          + " attempted to return null from a method with a primitive return type (" + method.getReturnType() + ").");
    }
    return result;
  }

　　（3）上面的案例是 select 语句，返回结果是List集合，所以进入 MapperMethod#executeForMany()：

  private <E> Object executeForMany(SqlSession sqlSession, Object[] args) {
    List<E> result;
    //获取参数
    Object param = method.convertArgsToSqlCommandParam(args);
    //是否有分页查询
    if (method.hasRowBounds()) {
      RowBounds rowBounds = method.extractRowBounds(args);
      result = sqlSession.<E>selectList(command.getName(), param, rowBounds);
    } else {
      result = sqlSession.<E>selectList(command.getName(), param);
    }
    // issue #510 Collections & arrays support
    //如果list中的泛型跟结果类型不一致，进行转换
    if (!method.getReturnType().isAssignableFrom(result.getClass())) {
      if (method.getReturnType().isArray()) {
        return convertToArray(result);
      } else {
        return convertToDeclaredCollection(sqlSession.getConfiguration(), result);
      }
    }
    return result;
  }

　　（4）selectList 执行了 DefaultSqlSession#selectList()：

  public <E> List<E> selectList(String statement, Object parameter) {
    return this.selectList(statement, parameter, RowBounds.DEFAULT);
  }

  public <E> List<E> selectList(String statement, Object parameter, RowBounds rowBounds) {
    try {
      //SQL执行的信息：resource（xxMapper.xml）、id、sql、返回类型等
      MappedStatement ms = configuration.getMappedStatement(statement);
      //执行查询
      return executor.query(ms, wrapCollection(parameter), rowBounds, Executor.NO_RESULT_HANDLER);
    } catch (Exception e) {
      throw ExceptionFactory.wrapException("Error querying database.  Cause: " + e, e);
    } finally {
      ErrorContext.instance().reset();
    }
  }

 

　　（5）接下来调用缓存执行器的方法：CachingExecutor#query()

  public <E> List<E> query(MappedStatement ms, Object parameterObject, RowBounds rowBounds, ResultHandler resultHandler) throws SQLException {
    //获取到执行SQL
    BoundSql boundSql = ms.getBoundSql(parameterObject);
    //将SQL包装成一个缓存对对象，该对象和结果集组成键值对存储到缓存中，方便下次直接从缓存中拿而不需要再次查询
    //createCacheKey：调用BaseExecutor#createCacheKey
    CacheKey key = createCacheKey(ms, parameterObject, rowBounds, boundSql);
    return query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
  }

  public <E> List<E> query(MappedStatement ms, Object parameterObject, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql)
      throws SQLException {
    //获取缓存
    Cache cache = ms.getCache();
    if (cache != null) {
      flushCacheIfRequired(ms);
      if (ms.isUseCache() && resultHandler == null) {
        ensureNoOutParams(ms, boundSql);
        @SuppressWarnings("unchecked")
        List<E> list = (List<E>) tcm.getObject(cache, key);
        if (list == null) {
          list = delegate.<E> query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
          tcm.putObject(cache, key, list); // issue #578 and #116
        }
        return list;
      }
    }
    //没有缓存连接查询
    return delegate.<E> query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
  }

　　（6）接下来执行 BaseExecutor#query()：从下面可以看到将结果缓存到 localCache 中了

  public <E> List<E> query(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql) throws SQLException {
    ErrorContext.instance().resource(ms.getResource()).activity("executing a query").object(ms.getId());
    if (closed) {
      throw new ExecutorException("Executor was closed.");
    }
    //如果不是嵌套查询（默认为0），且 <select> 的 flushCache=true 时清空缓存
    if (queryStack == 0 && ms.isFlushCacheRequired()) {
      clearLocalCache();
    }
    List<E> list;
    try {
     //嵌套查询层数+1
      queryStack++;
      //从localCache缓存中获取
      list = resultHandler == null ? (List<E>) localCache.getObject(key) : null;
      if (list != null) {
        handleLocallyCachedOutputParameters(ms, key, parameter, boundSql);
      } else {
        //连接查询
        list = queryFromDatabase(ms, parameter, rowBounds, resultHandler, key, boundSql);
      }
    } finally {
      queryStack--;
    }
    //下面是延迟加载逻辑
    if (queryStack == 0) {
      for (DeferredLoad deferredLoad : deferredLoads) {
        deferredLoad.load();
      }
      // issue #601
      deferredLoads.clear();
      if (configuration.getLocalCacheScope() == LocalCacheScope.STATEMENT) {
        // issue #482
        clearLocalCache();
      }
    }
    return list;
  }

  private <E> List<E> queryFromDatabase(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql) throws SQLException {
    List<E> list;
    //缓存中添加占位符
    localCache.putObject(key, EXECUTION_PLACEHOLDER);
    try {
      //连接查询获取到数据结果    
      list = doQuery(ms, parameter, rowBounds, resultHandler, boundSql);
    } finally {
      //删除占位符
      localCache.removeObject(key);
    }
    //将结果缓存起来
    localCache.putObject(key, list);
    //处理存储过程
    if (ms.getStatementType() == StatementType.CALLABLE) {
      localOutputParameterCache.putObject(key, parameter);
    }
    return list;
  }

　　2.但当 Spring Framework + Mybatis 时，情况就不一样了，每次查询都会连接数据库查询，控制台都会打印 SQL 出来，如下案例：

@Service
public class PersonService  {
    @Autowired
    PersonMapper personMapper;

    public List<Person> getList() {
        personMapper.list();
        personMapper.list();
        return personMapper.list();
    }
}

@Configuration
@ComponentScan("com.hrh")
@MapperScan("com.hrh.mapper")
public class MyBatisConfig {
    @Bean
    public SqlSessionFactoryBean sqlSessionFactory() throws Exception {
        SqlSessionFactoryBean factoryBean = new SqlSessionFactoryBean();
        factoryBean.setDataSource(dataSource());
        factoryBean.setMapperLocations(resolveMapperLocations());
        return factoryBean;
    }

    public Resource[] resolveMapperLocations() {
        ResourcePatternResolver resourceResolver = new PathMatchingResourcePatternResolver();
        List<String> mapperLocations = new ArrayList<>();
        mapperLocations.add("classpath*:com/hrh/mapper/*Mapper*.xml");
        List<Resource> resources = new ArrayList();
        if (mapperLocations != null) {
            for (String mapperLocation : mapperLocations) {
                try {
                    Resource[] mappers = resourceResolver.getResources(mapperLocation);
                    resources.addAll(Arrays.asList(mappers));
                } catch (IOException e) {
                    // ignore
                }
            }
        }
        return resources.toArray(new Resource[resources.size()]);
    }

    @Bean
    public DataSource dataSource() {
        DriverManagerDataSource driverManagerDataSource = new DriverManagerDataSource();
        driverManagerDataSource.setDriverClassName("com.mysql.jdbc.Driver");
        driverManagerDataSource.setUsername("xxx");
        driverManagerDataSource.setPassword("xxx");
        driverManagerDataSource.setUrl("jdbc:mysql://localhost:3306/xxx?useUnicode=true&characterEncoding=utf-8&autoReconnect=true");
        return driverManagerDataSource;
    }
}

        AnnotationConfigApplicationContext context = new AnnotationConfigApplicationContext(MyBatisConfig.class);
        PersonService bean = context.getBean(PersonService.class);
        bean.getList();

 

　　下面debug进入的步骤跟上面的（1）、（2）、（3）是一致的，但第四步却是进入 SqlSessionTemplate#selectList() 中【SqlSessionTemplate是mybatis-spring-xx.jar的，上文的DefaultSqlSession是属于mybatis-xx.jar的】：

  public <E> List<E> selectList(String statement, Object parameter) {
    return this.selectList(statement, parameter, RowBounds.DEFAULT);
  }

　　接下来的 selectList() 会被方法拦截：method.invoke() 会执行到 DefaultSqlSession#selectList()，重新回到上文的第四步并且继续下去，也就是在上文的（1）~（6）中插入了前后文，在其中做了关闭会话的操作；

  private class SqlSessionInterceptor implements InvocationHandler {
    @Override
    public Object invoke(Object proxy, Method method, Object[] args) throws Throwable {
      //得到会话
      SqlSession sqlSession = getSqlSession(
          SqlSessionTemplate.this.sqlSessionFactory,
          SqlSessionTemplate.this.executorType,
          SqlSessionTemplate.this.exceptionTranslator);
      try {
        //执行方法查询
        Object result = method.invoke(sqlSession, args);
        if (!isSqlSessionTransactional(sqlSession, SqlSessionTemplate.this.sqlSessionFactory)) {
          // force commit even on non-dirty sessions because some databases require
          // a commit/rollback before calling close()
          sqlSession.commit(true);//在关闭会话前提交和回滚
        }
        return result;
        
      } catch (Throwable t) {//有异常抛出异常并结束会话
        Throwable unwrapped = unwrapThrowable(t);
        if (SqlSessionTemplate.this.exceptionTranslator != null && unwrapped instanceof PersistenceException) {
          // release the connection to avoid a deadlock if the translator is no loaded. See issue #22
          closeSqlSession(sqlSession, SqlSessionTemplate.this.sqlSessionFactory);
          sqlSession = null;
          Throwable translated = SqlSessionTemplate.this.exceptionTranslator.translateExceptionIfPossible((PersistenceException) unwrapped);
          if (translated != null) {
            unwrapped = translated;
          }
        }
        throw unwrapped;
      } finally {
        //关闭会话
        if (sqlSession != null) {
          closeSqlSession(sqlSession, SqlSessionTemplate.this.sqlSessionFactory);
        }
      }
    }
  }

 

总结：

　　Mybatis 的一级缓存是会话级别的缓存（单线程的，特别鸡肋），Mybatis 每创建一个 SqlSession 会话对象，就表示打开一次数据库会话，在一次会话中，应用程序很可能在短时间内反复执行相同的查询语句，如果不对数据进行缓存，则每查询一次就要执行一次数据库查询，这就造成数据库资源的浪费。又因为通过 SqlSession 执行的操作，实际上由 Executor 来完成数据库操作的，所以在 Executor 中会建立一个简单的缓存，即一级缓存；将每次的查询结果缓存起来，再次执行查询的时候，会先查询一级缓存（默认开启的），如果命中，则直接返回，否则再去查询数据库并放入缓存中。

　　一级缓存的生命周期与 SqlSession 的生命周期相同，因此当 Mybatis 和 Spring Framework 的集成包中扩展了一个 SqlSessionTemplate 类（它是一个代理类，增强了查询方法），所有的查询经过 SqlSessionTemplate 代理拦截后再进入到 DefaultSqlSession#selectList() 中，结束查询后把会话SqlSession 关了，所以导致了缓存失效。

　　那为什么要这么操作呢？

　　原始的 Mybatis 有暴露 SqlSession 接口，因此有 close 方法暴露出来供你选择使用，你可以选择关与不关，但在 Mybatis 和 Spring Framework 的集成包中，SqlSession 是交给了 Spring Framework 管理的，没有暴露出来，为了稳妥决定，直接给你关了。