shardingjdbc 学习(四)-SQL路由实现

一 序

上一篇整理了jdbc的过程 ，update大概分两步，route跟执行。查询还有结果集归并。

上面本篇继续从路由开始整理。大概分为：分库分表整体，route包源码分析。

本文主要基于shardingjdbc 2.0.3版本整理。感谢芋道源码：http://www.iocoder.cn/Sharding-JDBC/sql-route-2/

二 分库分表：

这一部分算是第二篇API的补充吧。假设你阅读本文的时候已经熟悉相关概念或者阅读过官网的手册。

2.1 tablerule: rule包下面

表规则配置对象

public TableRule(final String logicTable, final List actualDataNodes, final Map dataSourceMap,
                     final ShardingStrategy databaseShardingStrategy, final ShardingStrategy tableShardingStrategy, 
                     final String generateKeyColumn, final KeyGenerator keyGenerator, final String logicIndex) {
        this.logicTable = logicTable.toLowerCase();
        this.actualDataNodes = null == actualDataNodes || actualDataNodes.isEmpty() ? generateDataNodes(logicTable, dataSourceMap) : generateDataNodes(actualDataNodes, dataSourceMap);
        this.databaseShardingStrategy = databaseShardingStrategy;
        this.tableShardingStrategy = tableShardingStrategy;
        this.generateKeyColumn = generateKeyColumn;
        this.keyGenerator = keyGenerator;
        this.logicIndex = null == logicIndex ? null : logicIndex.toLowerCase();
    }

有TableRuleConfiguration.build生成tablerule

2.2 DataNode 分库分表数据单元

 private List generateDataNodes(final String logicTable, final Map dataSourceMap) {
        List result = new LinkedList<>();
        for (String each : dataSourceMap.keySet()) {
            result.add(new DataNode(each, logicTable));//逻辑表
        }
        return result;
    }
    //物理表
    private List generateDataNodes(final List actualDataNodes, final Map dataSourceMap) {
        List result = new LinkedList<>();
        for (String each : actualDataNodes) {
            DataNode dataNode = new DataNode(each);
            Preconditions.checkArgument(dataSourceMap.containsKey(dataNode.getDataSourceName()), String.format("Cannot find data source in sharding rule, invalid actual data node is: '%s'", each));
            result.add(dataNode);
        }
        return result;
    }

    public DataNode(final String dataNode) {
        Preconditions.checkArgument(DataNode.isValidDataNode(dataNode), String.format("Invalid format for actual data nodes: '%s'", dataNode));
        List segments = Splitter.on(DELIMITER).splitToList(dataNode);
        dataSourceName = segments.get(0);
        tableName = segments.get(1);
    }

datanode 配置为 ${dataSourceName}.${tableName} 时。获取根据“.”截取数据源名称，表名。

2.3 shardingrule

分库分表规则配置对象,包含如下属性：

public final class ShardingRule {
    
    private final Map dataSourceMap;
    
    private final String defaultDataSourceName;
    
    private final Collection tableRules;
    
    private final Collection bindingTableRules = new LinkedList<>();
    
    private final ShardingStrategy defaultDatabaseShardingStrategy;
    
    private final ShardingStrategy defaultTableShardingStrategy;
    
    private final KeyGenerator defaultKeyGenerator;

dataSourceMap 数据源
tableRules 分表规则
BindingTableRule Binding表规则配置对象.
defaultDatabaseShardingStrategy 默认分库策略
defaultTableShardingStrategy 默认分表策略

defaultKeyGenerator 主键生成

   this.defaultDatabaseShardingStrategy = null == defaultDatabaseShardingStrategy ? new NoneShardingStrategy() : defaultDatabaseShardingStrategy;
        this.defaultTableShardingStrategy = null == defaultTableShardingStrategy ? new NoneShardingStrategy() : defaultTableShardingStrategy;

构造方法里面，没有分库分表策略，则 使用 NoneShardingStrategy 

三 SQL路由

整体源码包结构

3.1 SQLRouteResult

经过 SQL解析、SQL路由后，产生SQL路由结果，即 SQLRouteResult。根据路由结果，生成SQL，执行SQL。

public final class SQLRouteResult {
    
    private final SQLStatement sqlStatement;
    
    private final Set executionUnits = new LinkedHashSet<>();
    
    private final List generatedKeys = new LinkedList<>();
}

sqlStatement ：SQL语句对象，经过SQL解析的结果对象。
executionUnits ：SQL最小执行单元集合。SQL执行时，执行每个单元。

generatedKeys ：插入SQL语句生成的主键编号集合

3.2 分库分表策略

举个例子：StandardShardingStrategy

 @SuppressWarnings("unchecked")
    private Collection doSharding(final Collection availableTargetNames, final ListShardingValue shardingValue) {
        Collection result = new LinkedList<>();
        for (PreciseShardingValue each : transferToPreciseShardingValues(shardingValue)) {
            result.add(preciseShardingAlgorithm.doSharding(availableTargetNames, each));
        }
        return result;
    }

可见具体的实现有算法的dosharding来实现。shardingjdbc支持的具体的算法如下：

3.3 路由流程

本图来自与芋道源码整理

    public SQLRouteResult route(final List