MyBatis源码分析：SQL执行过程分析

Posted 2022-04-22 magic-sea

一、获取Mapper接口的代理

根据上一节，Mybatis初始化之后，利用sqlSession(defaultSqlSession)的getMapper方法获取Mapper接口

1 @Override
2 public <T> T getMapper(Class<T> type) 
3     return configuration.<T>getMapper(type, this);
4 

而调用configuration对象的getMapper方法

1 public <T> T getMapper(Class<T> type, SqlSession sqlSession) 
2     return mapperRegistry.getMapper(type, sqlSession);
3 

再次调用mapperRegister，注册mapper的类

 1 public <T> T getMapper(Class<T> type, SqlSession sqlSession) 
 2     final MapperProxyFactory<T> mapperProxyFactory = (MapperProxyFactory<T>) knownMappers.get(type);
 3     if (mapperProxyFactory == null) 
 4       throw new BindingException("Type " + type + " is not known to the MapperRegistry.");
 5     
 6     try 
 7       return mapperProxyFactory.newInstance(sqlSession);
 8      catch (Exception e) 
 9       throw new BindingException("Error getting mapper instance. Cause: " + e, e);
10     
11   

而mapperRegister根据传进来的mapper接口来创建MapperProxyFactory代理工厂对象，再用sqlSession参数创建Mapper的代理对象，这里运用的是JDK的动态代理，Proxy.newProxyInstance方法绑定mapper接口，第一个参数是类加载器，第二个参数是需要实现的接口数组，第三个是InvocationHandler接口，也就是交由InvocationHandler接口实现类MapperProxy里的invoke()方法去处理

1 @SuppressWarnings("unchecked")
2   protected T newInstance(MapperProxy<T> mapperProxy) 
3     return (T) Proxy.newProxyInstance(mapperInterface.getClassLoader(), new Class[]  mapperInterface , mapperProxy);
4   
5 
6   public T newInstance(SqlSession sqlSession) 
7     final MapperProxy<T> mapperProxy = new MapperProxy<>(sqlSession, mapperInterface, methodCache);
8     return newInstance(mapperProxy);
9   

然后就这样给UserMapper赋予了一个代理对象

1 UserMapper userMapper = sqlSession.getMapper(UserMapper.class);

二、使用Mapper代理对象进行查询操作

主代码调用代理对象查询，方法里面的参数为数据库字段的长整型id

1 user = userMapper.getUser(30L);

对应的mapper映射文件：

1 <select id="getUser" parameterType="long" resultMap="userMap">
2       SELECT user_id as id, user_name as username, sex, user_password as password, email from tb_user WHERE user_id = #id
3 </select>

使用Mapper代理对象，首先调用的是MapperProxy里面的invoke方法，传进三个主要的参数，分别是：代理对象、被调用的方法、方法的参数

 1 @Override
 2 public Object invoke(Object proxy, Method method, Object[] args) throws Throwable 
 3     try 
 4       if (Object.class.equals(method.getDeclaringClass())) 
 5         return method.invoke(this, args);
 6        else if (isDefaultMethod(method)) 
 7         return invokeDefaultMethod(proxy, method, args);
 8       
 9      catch (Throwable t) 
10       throw ExceptionUtil.unwrapThrowable(t);
11     
12     final MapperMethod mapperMethod = cachedMapperMethod(method);
13     return mapperMethod.execute(sqlSession, args);
14 

上面这段代码首先检查当前这个method是哪个类的方法，然后再判断有无默认方法，如果都没有则对方法进行缓存，最后对 SqlSession 进行的包装调用。

MapperMethod对SqlSession的操作进行了封装，来看其中的一段execute方法源码

 1 public Object execute(SqlSession sqlSession, Object[] args) 
 2     Object result;
 3     switch (command.getType()) 
 4       case INSERT: 
 5         Object param = method.convertArgsToSqlCommandParam(args);
 6         result = rowCountResult(sqlSession.insert(command.getName(), param));
 7         break;
 8       
 9       case UPDATE: 
10         Object param = method.convertArgsToSqlCommandParam(args);
11         result = rowCountResult(sqlSession.update(command.getName(), param));
12         break;
13       
14       case DELETE: 
15         Object param = method.convertArgsToSqlCommandParam(args);
16         result = rowCountResult(sqlSession.delete(command.getName(), param));
17         break;
18       
19       case SELECT:
20         if (method.returnsVoid() && method.hasResultHandler()) 
21           executeWithResultHandler(sqlSession, args);
22           result = null;
23          else if (method.returnsMany()) 
24           result = executeForMany(sqlSession, args);
25          else if (method.returnsMap()) 
26           result = executeForMap(sqlSession, args);
27          else if (method.returnsCursor()) 
28           result = executeForCursor(sqlSession, args);
29          else 
30           Object param = method.convertArgsToSqlCommandParam(args);
31           result = sqlSession.selectOne(command.getName(), param);
32           if (method.returnsOptional() &&
33               (result == null || !method.getReturnType().equals(result.getClass()))) 
34             result = Optional.ofNullable(result);
35           
36         
37         break;
38       case FLUSH:
39         result = sqlSession.flushStatements();
40         break;
41       default:
42         throw new BindingException("Unknown execution method for: " + command.getName());
43     
44     if (result == null && method.getReturnType().isPrimitive() && !method.returnsVoid()) 
45       throw new BindingException("Mapper method ‘" + command.getName()
46           + " attempted to return null from a method with a primitive return type (" + method.getReturnType() + ").");
47     
48     return result;
49 

调用的mapper的查询操作，先看看上面这段的SELECT这一段代码。首先是看方法的返回值类型是否为空并且结果处理器resultHandler，有的话则执行实现的ResultHandler的方法；之后也是检查方法的参数和返回类型，有的话执行各种情况下的方法；都没有的话，把参数传进SQL命令中

1 public Object convertArgsToSqlCommandParam(Object[] args) 
2       return paramNameResolver.getNamedParams(args);
3 

可以看到参数传递利用了ParamNameResolver，处理接口形式的参数，最后会把参数处放在一个map中，

 1 public Object getNamedParams(Object[] args) 
 2     final int paramCount = names.size();
 3     if (args == null || paramCount == 0) 
 4       return null;
 5      else if (!hasParamAnnotation && paramCount == 1) 
 6       return args[names.firstKey()];
 7      else 
 8       final Map<String, Object> param = new ParamMap<>();
 9       int i = 0;
10       for (Map.Entry<Integer, String> entry : names.entrySet()) 
11         param.put(entry.getValue(), args[entry.getKey()]);
12         // add generic param names (param1, param2, ...)
13         final String genericParamName = GENERIC_NAME_PREFIX + String.valueOf(i + 1);
14         // ensure not to overwrite parameter named with @Param
15         if (!names.containsValue(genericParamName)) 
16           param.put(genericParamName, args[entry.getKey()]);
17         
18         i++;
19       
20       return param;
21     
22  

参数解析完后，MapperMethod使用sqlSession，执行一条操作：

1 result = sqlSession.selectOne(command.getName(), param);

 1 @Override
 2 public <T> T selectOne(String statement, Object parameter) 
 3     // Popular vote was to return null on 0 results and throw exception on too many.
 4     List<T> list = this.selectList(statement, parameter);
 5     if (list.size() == 1) 
 6       return list.get(0);
 7      else if (list.size() > 1) 
 8       throw new TooManyResultsException("Expected one result (or null) to be returned by selectOne(), but found: " + list.size());
 9      else 
10       return null;
11     
12 

 1 @Override
 2 public <E> List<E> selectList(String statement, Object parameter, RowBounds rowBounds) 
 3     try 
 4       MappedStatement ms = configuration.getMappedStatement(statement);
 5       return executor.query(ms, wrapCollection(parameter), rowBounds, Executor.NO_RESULT_HANDLER);
 6      catch (Exception e) 
 7       throw ExceptionFactory.wrapException("Error querying database.  Cause: " + e, e);
 8      finally 
 9       ErrorContext.instance().reset();
10     
11 

sqlSession的selectList最后使用到MappedStatement，这个MappedStatement是保存Mapper中一个SQL语句的结点。利用执行器进行查询，第二个参数是为了检查参数是不是一个集合；

 1 private Object wrapCollection(final Object object) 
 2     if (object instanceof Collection) 
 3       StrictMap<Object> map = new StrictMap<>();
 4       map.put("collection", object);
 5       if (object instanceof List) 
 6         map.put("list", object);
 7       
 8       return map;
 9      else if (object != null && object.getClass().isArray()) 
10       StrictMap<Object> map = new StrictMap<>();
11       map.put("array", object);
12       return map;
13     
14     return object;
15 

第三个参数是rowBounds逻辑分页方式，这里使用的是默认的；第四个是执行器的参数，这里是null。

然后跳到了CacheExecutor的query方法，它根据传进的MappedStatement参数获取BoundSql对象，ms中有mapper中的sql语句，放在SqlSource，然后根据传进来的参数组装成boundSql；之后生成一个对应二级缓存的key，

1 @Override
2 public <E> List<E> query(MappedStatement ms, Object parameterObject, RowBounds rowBounds, ResultHandler resultHandler) throws SQLException 
3     BoundSql boundSql = ms.getBoundSql(parameterObject);
4     CacheKey key = createCacheKey(ms, parameterObject, rowBounds, boundSql);
5     return query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
6 

但是Mybatis默认只开启了一级缓存，本例中并没有开启二级缓存，所以直接执行最后一个父类delegate.query方法，

 1 public <E> List<E> query(MappedStatement ms, Object parameterObject, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql)
 2       throws SQLException 
 3     Cache cache = ms.getCache();
 4     if (cache != null) 
 5       flushCacheIfRequired(ms);
 6       if (ms.isUseCache() && resultHandler == null) 
 7         ensureNoOutParams(ms, boundSql);
 8         @SuppressWarnings("unchecked")
 9         List<E> list = (List<E>) tcm.getObject(cache, key);
10         if (list == null) 
11           list = delegate.query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
12           tcm.putObject(cache, key, list); // issue #578 and #116
13         
14         return list;
15       
16     
17     return delegate.query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
18 

上面调用的是BaseExecutor中的query方法，此方法中的最重要的一段代码

1       list = resultHandler == null ? (List<E>) localCache.getObject(key) : null;
2       if (list != null) 
3         handleLocallyCachedOutputParameters(ms, key, parameter, boundSql);
4        else 
5         list = queryFromDatabase(ms, parameter, rowBounds, resultHandler, key, 
6         boundSql);
7         

因为我没有自己写的resultHandler类，所以直接执行

1 　　　　list = queryFromDatabase(ms, parameter, rowBounds, resultHandler, key, boundSql);

其方法源码为BaseExecutor抽象类中的queryFromDataBase

 1 private <E> List<E> queryFromDatabase(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql) throws SQLException 
 2     List<E> list;
 3     localCache.putObject(key, EXECUTION_PLACEHOLDER);
 4     try 
 5       list = doQuery(ms, parameter, rowBounds, resultHandler, boundSql);
 6      finally 
 7       localCache.removeObject(key);
 8     
 9     localCache.putObject(key, list);
10     if (ms.getStatementType() == StatementType.CALLABLE) 
11       localOutputParameterCache.putObject(key, parameter);
12     
13     return list;
14 

queryFromDataBase中在深入，主要是第5行的doQuery方法：

 1 @Override
 2 public <E> List<E> doQuery(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler, BoundSql boundSql) throws SQLException 
 3     Statement stmt = null;
 4     try 
 5       Configuration configuration = ms.getConfiguration();
 6       StatementHandler handler = configuration.newStatementHandler(wrapper, ms, parameter, rowBounds, resultHandler, boundSql);
 7       stmt = prepareStatement(handler, ms.getStatementLog());
 8       return handler.query(stmt, resultHandler);
 9      finally 
10       closeStatement(stmt);
11     
12 

StatementHadler是四大核心对象之一，它的任务就是和数据库对话。上面这段代码configuration.newStatementHandler方法使用了RoutingStatementHandler(采用的适配器模式)创建StatementHandler：

 1 public RoutingStatementHandler(Executor executor, MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler, BoundSql boundSql) 
 2 
 3     switch (ms.getStatementType()) 
 4       case STATEMENT:
 5         delegate = new SimpleStatementHandler(executor, ms, parameter, rowBounds, resultHandler, boundSql);
 6         break;
 7       case PREPARED:
 8         delegate = new PreparedStatementHandler(executor, ms, parameter, rowBounds, resultHandler, boundSql);
 9         break;
10       case CALLABLE:
11         delegate = new CallableStatementHandler(executor, ms, parameter, rowBounds, resultHandler, boundSql);
12         break;
13       default:
14         throw new ExecutorException("Unknown statement type: " + ms.getStatementType());
15     
16 
17 

RoutingStatementHandler执行query方法

1 @Override
2 public <E> List<E> query(Statement statement, ResultHandler resultHandler) throws SQLException 
3     return delegate.<E>query(statement, resultHandler);
4 

PreparedStatementHandler执行query方法

1 @Override
2 public <E> List<E> query(Statement statement, ResultHandler resultHandler) throws SQLException 
3     PreparedStatement ps = (PreparedStatement) statement;
4     ps.execute();
5     return resultSetHandler.handleResultSets(ps);
6  

DefaultResultSetHandler执行handleResultSets方法，getFirstResultSet获取第一个结果集在于知道sql语句要操作到哪些元素数据（表的列），会获取到元数据名称、Java数据类型、JDBC数据类型，之后getResultMaps获取执行的sql配置的resultMap，之后一个resultMap对应一个结果集，依次遍历resultMap并处理结果集

 1 @Override
 2 public List<Object> handleResultSets(Statement stmt) throws SQLException 
 3     ErrorContext.instance().activity("handling results").object(mappedStatement.getId());
 4 
 5     final List<Object> multipleResults = new ArrayList<>();
 6 
 7     int resultSetCount = 0;
 8     ResultSetWrapper rsw = getFirstResultSet(stmt);
 9 
10     List<ResultMap> resultMaps = mappedStatement.getResultMaps();
11     int resultMapCount = resultMaps.size();
12     validateResultMapsCount(rsw, resultMapCount);
13     while (rsw != null && resultMapCount > resultSetCount)  //一个resultMap对应一个结果集，依次遍历resultMap并处理结果集
14       ResultMap resultMap = resultMaps.get(resultSetCount);
15       handleResultSet(rsw, resultMap, multipleResults, null); // 处理结果集
16       rsw = getNextResultSet(stmt);// 获取下一个结果集
17       cleanUpAfterHandlingResultSet();
18       resultSetCount++;
19     
20 
21     String[] resultSets = mappedStatement.getResultSets();
22     if (resultSets != null) 
23       while (rsw != null && resultSetCount < resultSets.length) 
24         ResultMapping parentMapping = nextResultMaps.get(resultSets[resultSetCount]);
25         if (parentMapping != null) 
26           String nestedResultMapId = parentMapping.getNestedResultMapId();
27           ResultMap resultMap = configuration.getResultMap(nestedResultMapId);
28           handleResultSet(rsw, resultMap, null, parentMapping);
29         
30         rsw = getNextResultSet(stmt);
31         cleanUpAfterHandlingResultSet();
32         resultSetCount++;
33       
34     
35 
36     return collapseSingleResultList(multipleResults); //把结果集转化为List
37 

DefaultResultSetHandler的getFirstResultSet方法

 1 private ResultSetWrapper getFirstResultSet(Statement stmt) throws SQLException 
 2     ResultSet rs = stmt.getResultSet();
 3     while (rs == null)  // 没有结果集，也许是数据库驱动还没有返回第一个结果集
 4       // move forward to get the first resultset in case the driver
 5       // doesn‘t return the resultset as the first result (HSQLDB 2.1)
 6       if (stmt.getMoreResults())  // 尝试再一次获取结果集
 7         rs = stmt.getResultSet();
 8        else 
 9         if (stmt.getUpdateCount() == -1)   //表示驱动已经返回，没有更多结果，没有结果集
10           // no more results. Must be no resultset
11           break;
12         
13       
14     
15     return rs != null ? new ResultSetWrapper(rs, configuration) : null; //不为空则返回结果集的包装
16 

ResultSetWrapper构造函数（包装结果集）

 1 public ResultSetWrapper(ResultSet rs, Configuration configuration) throws SQLException 
 2     super();
 3     this.typeHandlerRegistry = configuration.getTypeHandlerRegistry();
 4     this.resultSet = rs;
 5     final ResultSetMetaData metaData = rs.getMetaData();
 6     final int columnCount = metaData.getColumnCount();
 7     for (int i = 1; i <= columnCount; i++)  // 设置结果集的元数据
 8       columnNames.add(configuration.isUseColumnLabel() ? metaData.getColumnLabel(i) : metaData.getColumnName(i));
 9       jdbcTypes.add(JdbcType.forCode(metaData.getColumnType(i)));
10       classNames.add(metaData.getColumnClassName(i));
11     
12 

把结果集转化为List

1 @SuppressWarnings("unchecked")
2 private List<Object> collapseSingleResultList(List<Object> multipleResults) 
3     return multipleResults.size() == 1 ? (List<Object>) multipleResults.get(0) : multipleResults;
4 

然后一层层传递回去，最后获得查询结果。（待续）