Sharding-Jdbc源码学习（三）：shardingDataSource执行sql过程解析

前一篇博文中提到JdbcTeplate执行SQL的第一步首先是创建连接对象。

1，DataSourceUtils.getConnection（obtainDataSource（））; 创建连接，该处只是获得连接对象并未真正建立连接

2，由“ Connection ”创建对应的Connection.createStatement（）创建对应的Statement

3，执行Statement的execute（final String sql）方法

4，返回结果集

一，创建连接对象

 io.shardingjdbc.core.jdbc.core.datasource.ShardingDataSource.getConnection（）   源码如下

@Override
public ShardingConnection getConnection() throws SQLException {
    return new ShardingConnection(shardingContext);  --继承了Connection
}

二，调用ShardingConnection创建³³ 声明，源码如下

io.shardingjdbc.core.jdbc.core.connection.ShardingConnection.createStatement() 源码如下

@Override
public Statement createStatement() throws SQLException {
    return new ShardingStatement(this);
}

三，执行Statement的execute（final String sql）方法

 即执行io.shardingjdbc.core.jdbc.core.statement.ShardingStatement。执行（SQL）方法，源码如下

io.shardingjdbc.core.jdbc.core.statement.ShardingStatement.execute(sql)

@Override
public boolean execute(final String sql) throws SQLException {
    try {
        return generateExecutor(sql).execute();
    } finally {
        currentResultSet = null;
    }
}

由此可见执行SQL时需要调用对应执行器的execute（）方法，具体调用源码如下。

io.shardingjdbc.core.jdbc.core.statement.ShardingStatement.generateExecutor(sql)

private StatementExecutor generateExecutor(final String sql) throws SQLException {
    clearPrevious();
    routeResult = new StatementRoutingEngine(connection.getShardingContext()).route(sql);
    Collection statementUnits = new LinkedList<>();
    for (SQLExecutionUnit each : routeResult.getExecutionUnits()) {
        Collection connections;
        SQLType sqlType = routeResult.getSqlStatement().getType();
        if (SQLType.DDL == sqlType) {
            connections = connection.getConnectionsForDDL(each.getDataSource());
        } else {
            connections = Collections.singletonList(connection.getConnection(each.getDataSource(), routeResult.getSqlStatement().getType()));
        }
        for (Connection connection : connections) {
            Statement statement = connection.createStatement(resultSetType, resultSetConcurrency, resultSetHoldability);
            replayMethodsInvocation(statement);
            statementUnits.add(new StatementUnit(each, statement));
            routedStatements.add(statement);
        }
    }
    return new StatementExecutor(connection.getShardingContext().getExecutorEngine(), routeResult.getSqlStatement().getType(), statementUnits);
}

总结：上述源码重点描述了sharding-jdbc如何进行SQL解析，SQL改写，SQL路由。该代码拆分为如下3步子功能：

1、根据sql获取对应的sql路由结果，即获取该sql具体和哪个datasource建立链接

SQLRouteResult routeResult = new StatementRoutingEngine(connection.getShardingContext()).route(sql); 

 routeResult 即为sql路由结果，即获取该sql应该和哪个datasource建立连接？？

2、根绝sql路由结果和对应的datasource建立连接

Collection connections;
        SQLType sqlType = routeResult.getSqlStatement().getType();
        if (SQLType.DDL == sqlType) {
            connections = connection.getConnectionsForDDL(each.getDataSource());     真正的建立连接
        } else {
            connections = Collections.singletonList(connection.getConnection(each.getDataSource(), routeResult.getSqlStatement().getType()));   真正的建立连接

   }

3、sql执行器执行对应的sql