MyBatis框架原理3：缓存

上一篇[MyBatis框架原理2：SqlSession运行过程][1]介绍了MyBatis的工作流程，其中涉及到了MyBatis缓存的使用，首先回顾一下工作流程图：

如果开启了二级缓存，数据查询执行过程就是首先从二级缓存中查询，如果未命中则从一级缓存中查询，如果也未命中则从数据库中查询。MyBatis的一级和二级缓存都是基于Cache接口的实现，下面先来看看Cache接口和其各种实现类。

Cache接口及常用装饰器

public interface Cache {
  String getId();
  //缓存中添加数据，key为生成的CacheKey,value为查询结果
  void putObject(Object key, Object value);
  //查询
  Object getObject(Object key);
  //删除
  Object removeObject(Object key);
  //清空缓存
  void clear();
  //获取缓存数量
  int getSize();
  //获取读写锁
  ReadWriteLock getReadWriteLock();
}

Cache接口位于MyBatis的cache包下，定义了缓存的基本方法，其实现类采用了装饰器模式，通过实现类的组装，可以实现操控缓存的功能。cache包结构如下：

• PerpetualCache是Cache接口的实现类，通过内部的HashMap来对缓存进行基本的操作，通常配合装饰器类一起使用。
• BlockingCache装饰器：保证只有一个线程到数据库中查询指定key的数据，如果该线程在BlockingCache中未查找到数据，就获取key对应的锁，阻塞其他查询这个key的线程，通过其内部ConcurrentHashMap来实现，源码如下：

public class BlockingCache implements Cache {
  //阻塞时长
  private long timeout;
  private final Cache delegate;
  //key和ReentrantLock对象一一对应
  private final ConcurrentHashMap<Object, ReentrantLock> locks;
  @Override
  public Object getObject(Object key) {
    //获取key的锁
    acquireLock(key);
    //根据key查询
    Object value = delegate.getObject(key);
    //如果命中缓存，释放锁，未命中则继续持有锁
    if (value != null) {
      releaseLock(key);
    }
    return value;
  }
 @Override
  //从数据库获取结果后，将结果放入BlockingCache,然后释放锁
  public void putObject(Object key, Object value) {
    try {
      delegate.putObject(key, value);
    } finally {
      releaseLock(key);
    }
  }
...

• FifoCache装饰器: 先入先出规则删除最早的缓存，通过其内部的Deque实现。
• LruCache装饰器: 删除最近使用最少的缓存， 通过内部的LinkedHashMap实现。
• SynchronizedCache装饰器：同步Cache。
• LoggingCache装饰器： 提供日志功能，记录和输出缓存命中率。
• SerializedCache装饰器：序列化功能。

CacheKey

CacheKey对象是用来确认缓存项的唯一标识，由其内部ArrayList添加的所有对象来确认两个CacheKey是否相同，通常ArrayList内将添加MappedStatement的id,SQL语句，用户传递给SQL语句的参数以及查询结果集范围RowBounds等，CacheKey源码如下:

public class CacheKey implements Cloneable, Serializable {
...
  private final int multiplier;
  private int hashcode;
  private long checksum;
  private int count;
  private List<Object> updateList;
  public CacheKey() {
    this.hashcode = DEFAULT_HASHCODE;
    this.multiplier = DEFAULT_MULTIPLYER;
    this.count = 0;
    this.updateList = new ArrayList<Object>();
  }
  //向updateLis中添加对象
  public void update(Object object) {
    int baseHashCode = object == null ? 1 : ArrayUtil.hashCode(object); 
    count++;
    checksum += baseHashCode;
    baseHashCode *= count;
    hashcode = multiplier * hashcode + baseHashCode;
    updateList.add(object);
  }
  @Override
  //重写equals方法判断CacheKey是否相同
  public boolean equals(Object object) {
    if (this == object) {
      return true;
    }
    if (!(object instanceof CacheKey)) {
      return false;
    }
    final CacheKey cacheKey = (CacheKey) object;
    if (hashcode != cacheKey.hashcode) {
      return false;
    }
    if (checksum != cacheKey.checksum) {
      return false;
    }
    if (count != cacheKey.count) {
      return false;
    }
    //比较updateList中每一项
    for (int i = 0; i < updateList.size(); i++) {
      Object thisObject = updateList.get(i);
      Object thatObject = cacheKey.updateList.get(i);
      if (!ArrayUtil.equals(thisObject, thatObject)) {
        return false;
      }
    }
    return true;
  }
}

一级缓存

一级缓存是session级别缓存，只存在当前会话中，在没有任何配置下，MyBatis默认开启一级缓存，当一个SqlSession第一次执行SQL语句和参数查询时，将生成的CacheKey和查询结果放入缓存中，下一次通过相同的SQL语句和参数查询时，就会从缓存中获取，当进行更新或者插入操作时，一级缓存会进行清空。在上一篇中说到，MayBatis进行一级缓存查询和写入是由BaseExecutor执行的，源码如下：

• 初始化缓存：
  
  一级缓存是Cache接口的PerpetualCache实现类对象

public abstract class BaseExecutor implements Executor {
  ...
  protected PerpetualCache localCache;
  protected PerpetualCache localOutputParameterCache;
  protected Configuration configuration;
  protected int queryStack;
  private boolean closed;
  protected BaseExecutor(Configuration configuration, Transaction transaction) {
    this.transaction = transaction;
    this.deferredLoads = new ConcurrentLinkedQueue<DeferredLoad>();
    //一级缓存初始化
    this.localCache = new PerpetualCache("LocalCache");
    this.localOutputParameterCache = new PerpetualCache("LocalOutputParameterCache");
    this.closed = false;
    this.configuration = configuration;
    this.wrapper = this;
  }
...

• 生成CacheKey
  
  BaseExecutor生成CacheKey，CacheKey的updateList中放入了MappedStatement,传入SQL的参数，结果集范围rowBounds和boundSql:

  @Override
  public <E> List<E> query(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler) throws SQLException {
    BoundSql boundSql = ms.getBoundSql(parameter);
    CacheKey key = createCacheKey(ms, parameter, rowBounds, boundSql);
    return query(ms, parameter, rowBounds, resultHandler, key, boundSql);
 }

• 将查询结果和CacheKey放入缓存：

private <E> List<E> queryFromDatabase(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql) throws SQLException {
    List<E> list;
    //缓存中放入CacheKey和占位符
    localCache.putObject(key, EXECUTION_PLACEHOLDER);
    try {
      //在数据库中查询操作
      list = doQuery(ms, parameter, rowBounds, resultHandler, boundSql);
    } finally {
      localCache.removeObject(key);
    }
    //缓存中放入CacheKey和结果集
    localCache.putObject(key, list);
    if (ms.getStatementType() == StatementType.CALLABLE) {
      localOutputParameterCache.putObject(key, parameter);
    }
    //返回结果
    return list;
  }

• 再次执行相同查询条件时从缓存获取结果：

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.");
    }
    if (queryStack == 0 && ms.isFlushCacheRequired()) {
      clearLocalCache();
    }
    List<E> list;
    try {
      queryStack++;
      //从缓存获取结果
      list = resultHandler == null ? (List<E>) localCache.getObject(key) : null;
      if (list != null) {
        handleLocallyCachedOutputParameters(ms, key, parameter, boundSql);
      } else {
      //未命中缓存，则从数据库查询
        list = queryFromDatabase(ms, parameter, rowBounds, resultHandler, key, boundSql);
      }
    } finally {
      queryStack--;
    }
    if (queryStack == 0) {
      for (DeferredLoad deferredLoad : deferredLoads) {
        deferredLoad.load();
      }
      // issue #601
      deferredLoads.clear();
      if (configuration.getLocalCacheScope() == LocalCacheScope.STATEMENT) {
        // issue #482
        clearLocalCache();
      }
    }
    return list;
  }

• 更新操作时清空缓存：

 public int update(MappedStatement ms, Object parameter) throws SQLException {
    ErrorContext.instance().resource(ms.getResource()).activity("executing an update").object(ms.getId());
    if (closed) {
      throw new ExecutorException("Executor was closed.");
    }
    //清空缓存
    clearLocalCache();
    return doUpdate(ms, parameter);
  }

通过以下代码验证下，分别开两个session进行相同的查询，第一个session查询两次：

public void testSelect() {
		SqlSession sqlSession = sqlSessionFactory.openSession();
		User user = sqlSession.selectOne("findUserById", 1);
		System.out.println(user);
		User user2 = sqlSession.selectOne("findUserById", 1);
		System.out.println(user2);
		sqlSession.close();
		System.out.println("sqlSession closed!===================================");
		//新建会话
		SqlSession sqlSession2 = sqlSessionFactory.openSession();
		User user3 = sqlSession2.selectOne("findUserById", 1);
		System.out.println(user3);
		sqlSession2.close();
	}

把日志设置为DEBUG级别得到运行日志：

DEBUG [main] - ==>  Preparing: SELECT * FROM user WHERE id = ?
DEBUG [main] - ==> Parameters: 1(Integer)
DEBUG [main] - <==      Total: 1
User [id=1, username=小明, birthday=null, sex=1, address=四川成都]
User [id=1, username=小明, birthday=null, sex=1, address=四川成都]
DEBUG [main] - Resetting autocommit to true on JDBC Connection [com.mysql.jdbc.JDBC4Connection@16022d9d]
DEBUG [main] - Closing JDBC Connection [com.mysql.jdbc.JDBC4Connection@16022d9d]
DEBUG [main] - Returned connection 369241501 to pool.
sqlSession closed!===================================
DEBUG [main] - Opening JDBC Connection
DEBUG [main] - Checked out connection 369241501 from pool.
DEBUG [main] - Setting autocommit to false on JDBC Connection [com.mysql.jdbc.JDBC4Connection@16022d9d]
DEBUG [main] - ==>  Preparing: SELECT * FROM user WHERE id = ?
DEBUG [main] - ==> Parameters: 1(Integer)
DEBUG [main] - <==      Total: 1
User [id=1, username=小明, birthday=null, sex=1, address=四川成都]
DEBUG [main] - Resetting autocommit to true on JDBC Connection [com.mysql.jdbc.JDBC4Connection@16022d9d]
DEBUG [main] - Closing JDBC Connection [com.mysql.jdbc.JDBC4Connection@16022d9d]
DEBUG [main] - Returned connection 369241501 to pool.

第一次会话中，虽然查询了两次id为1的用户，但是只执行了一次SQL，关闭会话后开启一次新的会话，再次查询id为1的用户，SQL再次执行，说明了一级缓存只存在SqlSession中，不同SqlSession不能共享。

二级缓存

二级缓存是Mapper级别缓存，也就是同一Mapper下不同的session共享二级缓存区域。

只需要在XML映射文件中增加cache标签或cache-ref标签标签就可以开启二级缓存，cache-ref标签配置的是共享其指定Mapper的二级缓存区域。具体配置信息如下：

• blocking : 是否使用阻塞缓存
• readOnly : 是否只读
• eviction: 缓存策略，可指定Cache接口下装饰器类FifoCache、LruCache、SoftCache和WeakCache
• flushInterval : 自动刷新缓存时间
• size : 设置缓存个数
• type : 设置缓存类型，用于自定义缓存类，默认为PerpetualCache

二级缓存是在MyBatis的解析配置文件时初始化，在XMLMapperBuilder中将缓存配置解析：

private void cacheElement(XNode context) throws Exception {
    if (context != null) {
      //指定默认类型为PerpetualCache
      String type = context.getStringAttribute("type", "PERPETUAL");
      Class<? extends Cache> typeClass = typeAliasRegistry.resolveAlias(type);
      //默认缓存策略为LruCache
      String eviction = context.getStringAttribute("eviction", "LRU");
      Class<? extends Cache> evictionClass = typeAliasRegistry.resolveAlias(eviction);
      Long flushInterval = context.getLongAttribute("flushInterval");
      Integer size = context.getIntAttribute("size");
      boolean readWrite = !context.getBooleanAttribute("readOnly", false);
      boolean blocking = context.getBooleanAttribute("blocking", false);
      Properties props = context.getChildrenAsProperties();
      //委托builderAssistant构建二级缓存
      builderAssistant.useNewCache(typeClass, evictionClass, flushInterval, size, readWrite, blocking, props);
    }
  }

构建过程：

public Cache useNewCache(Class<? extends Cache> typeClass,
      Class<? extends Cache> evictionClass,
      Long flushInterval,
      Integer size,
      boolean readWrite,
      boolean blocking,
      Properties props) {
    Cache cache = new CacheBuilder(currentNamespace)
        //设置缓存类型，默认为PerpetualCache
        .implementation(valueOrDefault(typeClass, PerpetualCache.class))
        //设置缓存策略，默认使用LruCache装饰器
        .addDecorator(valueOrDefault(evictionClass, LruCache.class))
        //设置刷新时间
        .clearInterval(flushInterval)
        //设置大小
        .size(size)
        //设置是否只读
        .readWrite(readWrite)
        .blocking(blocking)
        .properties(props)
        .build();
    configuration.addCache(cache);
    currentCache = cache;
    return cache;
  }

最终得到默认的二级缓存对象结构为：

CachingExecutor将初始化的Cache对象用TransactionalCache包装后放入TransactionalCacheManager的Map中，下面代码中的tcm就是TransactionalCacheManager对象，CachingExecutor执行二级缓存操作过程：

public <E> List<E> query(MappedStatement ms, Object parameterObject, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql)
      throws SQLException {
    //从Configuration的MappedStatement中获取二级缓存
    Cache cache = ms.getCache();
    if (cache != null) {
     //判断是否需要刷新缓存，SELECT不刷新，INSERT|UPDATE|DELETE刷新缓存
      flushCacheIfRequired(ms);
      if (ms.isUseCache() && resultHandler == null) {
        ensureNoOutParams(ms, boundSql);
        @SuppressWarnings("unchecked")
        //从二级缓存中获取数据
        List<E> list = (List<E>) tcm.getObject(cache, key);
        if (list == null) {
          //委托BaseExecutor查询
          list = delegate.<E> query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
          //查询结果放入二级缓存
          tcm.putObject(cache, key, list); // issue #578 and #116
        }
        return list;
      }
    }
    return delegate.<E> query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
  }

通过之前一级缓存的例子验证二级缓存，只需要在UserMapper映射文件中加入cache标签，并且让相关POJO类实现java.io.Serializable接口，运行得到日志：

DEBUG [main] - ==>  Preparing: SELECT * FROM user WHERE id = ?
DEBUG [main] - ==> Parameters: 1(Integer)
DEBUG [main] - <==      Total: 1
User [id=1, username=小明, birthday=null, sex=1, address=四川成都]
DEBUG [main] - Cache Hit Ratio [com.kkb.mybatis.mapper.UserMapper]: 0.0
User [id=1, username=小明, birthday=null, sex=1, address=四川成都]
DEBUG [main] - Resetting autocommit to true on JDBC Connection [com.mysql.jdbc.JDBC4Connection@5c072e3f]
DEBUG [main] - Closing JDBC Connection [com.mysql.jdbc.JDBC4Connection@5c072e3f]
DEBUG [main] - Returned connection 1543974463 to pool.
sqlSession closed!===================================
DEBUG [main] - Cache Hit Ratio [com.kkb.mybatis.mapper.UserMapper]: 0.3333333333333333
User [id=1, username=小明, birthday=null, sex=1, address=四川成都]

不同session查询同一条记录时，总共只执行了一次SQL语句，并且日志打印出了缓存的命中率，这时候不同session已经共享了二级缓存区域。

[1]: https://www.cnblogs.com/abcboy/p/9656302.html