hint

配置文件

  1. spring:
  2.   shardingsphere:
  3.     datasource:
  4.       ds0:
  5.         type: com.alibaba.druid.pool.DruidDataSource
  6.         url: jdbc:mysql://localhost:3306/sharding?characterEncoding=utf-8
  7.         username: root
  8.         password: root
  9.       names: ds0
  10.     sharding:
  11.       tables:
  12.         x_user:
  13.           actual-data-nodes: ds0.x_user_$->{0..1}
  14.           table-strategy:
  15.             hint:
  16.               algorithmClassName: org.sharding.shardingsphere.TableHintShardingAlgorithm
  17.           databaseStrategy:
  18.             hint:
  19.               algorithmClassName: org.sharding.shardingsphere.DbHintShardingAlgorithm

HintShardingAlgorithm

algorithmClassName配置项指定的算法类要实现HintShardingAlgorithm接口。

  1. public class TableHintShardingAlgorithm implements HintShardingAlgorithm<Integer> {
  4.     public Collection<String> doSharding(Collection<String> availableTargetNames, HintShardingValue<Integer> shardingValue) {
  5.         return availableTargetNames;
  6.     }
  7. }

availableTargetNames表示可用的目标名称，对于数据库就是可用的数据库，对于表就是可用的表，是通过配置项actual-data-nodes配置的。shardingValue表示可用的分片值，通过HintManager配置的。

HintManager

在执行sql前需要通过HintManager指定分片值。

  1. public List list() {
  2.         HintManager.clear();
  3.         // HintManager API 工具类实例
  4.         HintManager hintManager = HintManager.getInstance();
  5.         // 直接指定对应具体的数据库
  6.         hintManager.addDatabaseShardingValue("x_user",0);
  7.         hintManager.addDatabaseShardingValue("x_user",1);
  8.         // 设置表的分片健
  9.         hintManager.addTableShardingValue("x_user" , 0);
  10.         hintManager.addTableShardingValue("x_user" , 1);
  11.         hintManager.addTableShardingValue("x_user" , 2);
  13.         return xUserMapper.select();
  14.     }

inline

配置项

  1. spring:
  2.   shardingsphere:
  3.     datasource:
  4.       ds0:
  5.         type: com.alibaba.druid.pool.DruidDataSource
  6.         url: jdbc:mysql://localhost:3306/sharding?characterEncoding=utf-8
  7.         username: root
  8.         password: root
  9.       names: ds0
  10.     sharding:
  11.         t_order:
  12.           actual-data-nodes: ds0.t_order_$->{0..2}
  13.           table-strategy:
  14.             inline:
  15.               sharding-column: order_id
  16.               algorithm-expression: t_order_$->{order_id % 3}

需要配置分片列和算法表达式。

standard

配置信息

  1. spring:
  2.   shardingsphere:
  3.     datasource:
  4.       ds0:
  5.         type: com.alibaba.druid.pool.DruidDataSource
  6.         url: jdbc:mysql://localhost:3306/sharding?characterEncoding=utf-8
  7.         username: root
  8.         password: root
  9.       names: ds0
  10.     sharding:
  11.       tables:
  12.         t_pay:
  13.           actual-data-nodes: ds0.t_pay_$->{0..2}
  14.           table-strategy:
  15.             standard:
  16.               sharding-column: pay_id
  17.               precise-algorithm-class-name: org.sharding.shardingsphere.TablePreciseShardingAlgorithm
  18.               range-algorithm-class-name: org.sharding.shardingsphere.TableRangeShardingAlgorithm

precise-algorithm-class-name：精确分片算法类名称，用于=和IN。。该类需实现PreciseShardingAlgorithm接口并提供无参数的构造器。range-algorithm-class-name：范围分片算法类名称，用于BETWEEN，可选。该类需实现RangeShardingAlgorithm接口并提供无参数的构造器。

PreciseShardingAlgorithm

  1. /**
  2.  * @Author: licheng
  3.  * @Date: 2021/7/3 20:02
  4.  */
  5. public class TablePreciseShardingAlgorithm implements PreciseShardingAlgorithm<Integer> {
  7.     public String doSharding(Collection<String> availableTargetNames, PreciseShardingValue<Integer> shardingValue) {
  8.         int l = shardingValue.getValue() % availableTargetNames.size();
  9.         String str = String.valueOf(l);
  10.         for (String tb : availableTargetNames) {
  11.             if (tb.endsWith(str)) {
  12.                 return tb;
  13.             }
  14.         }
  15.         return null;
  16.     }
  17. }

availableTargetNames表示可用目标，对于数据源分片表示可用的数据源，对于表分片表示可用的表，就是
actual-data-nodes配置的值。shardingValue表示分片值。

RangeShardingAlgorithm

  1. /**
  2.  * @Author: licheng
  3.  * @Date: 2021/7/3 20:07
  4.  */
  5. public class TableRangeShardingAlgorithm implements RangeShardingAlgorithm<Integer> {
  6.     public Collection<String> doSharding(Collection<String> availableTargetNames, RangeShardingValue<Integer> shardingValue) {
  8.         return availableTargetNames;
  9.     }
  10. }

RangeShardingValue表示范围分片值，availableTargetNames是可用目标。

complex

支持多分片键，提供对SQL语句中的=, IN和BETWEEN AND的分片操作支持。

配置项

  1. spring:
  2.   shardingsphere:
  3.     sharding:
  4.       tables:
  5.         t_msg:
  6.           actual-data-nodes: ds0.t_msg_$->{0..2}
  7.           table-strategy:
  8.             complex:
  9.               sharding-columns: src_user_id,dest_user_id
  10.               algorithm-class-name: org.sharding.shardingsphere.TableComplexKeysShardingAlgorithm

sharding-columns指定分片列，algorithm-class-name指定算法类，需要实现ComplexKeysShardingAlgorithm接口。

ComplexKeysShardingAlgorithm

ComplexKeysShardingAlgorithm是complex分片策略的分片算法类。

  1. /**
  2.  * @Author: licheng
  3.  * @Date: 2021/7/4 0:50
  4.  */
  5. public class TableComplexKeysShardingAlgorithm implements ComplexKeysShardingAlgorithm {
  6.     public Collection<String> doSharding(Collection availableTargetNames, ComplexKeysShardingValue shardingValue) {
  7.         System.out.println(shardingValue.getColumnNameAndRangeValuesMap());
  8.         System.out.println(shardingValue.getColumnNameAndShardingValuesMap());
  9.         return availableTargetNames;
  10.     }
  11. }

availableTargetNames：可用目标。ComplexKeysShardingValue：分片值，getColumnNameAndRangeValuesMap方法返回between的分片值，getColumnNameAndShardingValuesMap代表=和in的分片值。