Java API 形式配置的读写分离可以方便的适用于各种场景，不依赖额外的 jar 包，用户只需要通过 java 代码构造读写分离数据源便可以使用读写分离功能。
类名称：org.apache.shardingsphere.readwritesplitting.config.ReadwriteSplittingRuleConfiguration
可配置属性：
| 名称 | 数据类型 | 说明 |
|---|---|---|
| dataSources (+) | Collection<ReadwriteSplittingDataSourceRuleConfiguration> | 读写数据源配置 |
| loadBalancers (*) | Map<String, AlgorithmConfiguration> | 从库负载均衡算法配置 |
类名称：org.apache.shardingsphere.readwritesplitting.config.rule.ReadwriteSplittingDataSourceGroupRuleConfiguration
可配置属性：
| 名称 | 数据类型 | 说明 | 默认值 |
|---|---|---|---|
| name | String | 读写分离数据源名称 | - |
| writeDataSourceName | String | 写库数据源名称 | - |
| readDataSourceNames | List<String> | 读库数据源名称 | - |
| transactionalReadQueryStrategy (?) | TransactionalReadQueryStrategy | 事务内读请求的路由策略，可选值：PRIMARY（路由至主库）、FIXED（同一事务内路由至固定数据源）、DYNAMIC（同一事务内路由至非固定数据源） | DYNAMIC |
| loadBalancerName (?) | String | 读库负载均衡算法名称 | 轮询负载均衡算法 |
算法类型的详情，请参见内置负载均衡算法列表。
public DataSource getDataSource() throws SQLException {
ReadwriteSplittingDataSourceRuleConfiguration dataSourceConfig = new ReadwriteSplittingDataSourceRuleConfiguration(
"demo_read_query_ds", "demo_write_ds", Arrays.asList("demo_read_ds_0", "demo_read_ds_1"), "demo_weight_lb");
Properties algorithmProps = new Properties();
algorithmProps.setProperty("demo_read_ds_0", "2");
algorithmProps.setProperty("demo_read_ds_1", "1");
Map<String, AlgorithmConfiguration> algorithmConfigMap = new HashMap<>(1);
algorithmConfigMap.put("demo_weight_lb", new AlgorithmConfiguration("WEIGHT", algorithmProps));
ReadwriteSplittingRuleConfiguration ruleConfig = new ReadwriteSplittingRuleConfiguration(Collections.singleton(dataSourceConfig), algorithmConfigMap);
Properties props = new Properties();
props.setProperty("sql-show", Boolean.TRUE.toString());
return ShardingSphereDataSourceFactory.createDataSource(createDataSourceMap(), Collections.singleton(ruleConfig), props);
}
private Map<String, DataSource> createDataSourceMap() {
Map<String, DataSource> result = new HashMap<>(3, 1);
result.put("demo_write_ds", DataSourceUtil.createDataSource("demo_write_ds"));
result.put("demo_read_ds_0", DataSourceUtil.createDataSource("demo_read_ds_0"));
result.put("demo_read_ds_1", DataSourceUtil.createDataSource("demo_read_ds_1"));
return result;
}