shardingsphere-jdbc之inline分片(4.1.1)

PS:依赖配置看shardingsphere-jdbc第一篇

 数据表

CREATE TABLE `t_order0` (
  `order_id` bigint(20) NOT NULL AUTO_INCREMENT,
  `order_code` varchar(20) CHARACTER SET utf8mb4 COLLATE utf8mb4_bin NOT NULL DEFAULT '',
  `user_id` int(11) DEFAULT NULL,
  `amount` decimal(4,2) DEFAULT NULL,
  PRIMARY KEY (`order_id`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8mb4;

CREATE TABLE `t_order1` (
  `order_id` bigint(20) NOT NULL AUTO_INCREMENT,
  `order_code` varchar(20) CHARACTER SET utf8mb4 COLLATE utf8mb4_bin NOT NULL DEFAULT '',
  `user_id` int(11) DEFAULT NULL,
  `amount` decimal(4,2) DEFAULT NULL,
  PRIMARY KEY (`order_id`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8mb4;

yml

server:
  port: 8081
spring:
  main:
    allow-bean-definition-overriding: true
  shardingsphere:
    datasource:
      names: ds0
      ds0:
        type: com.alibaba.druid.pool.DruidDataSource
        driverClassName: "com.mysql.cj.jdbc.Driver"
        url: jdbc:mysql://xxxx:30037/xxx?characterEncoding=UTF-8&useSSL=false&zeroDateTimeBehavior=convertToNull&allowMultiQueries=true&serverTimezone=Asia/Shanghai
        username: "xxx"
        password: "xxx"
    sharding:
      tables:
        t_order:
          actualDataNodes: ds0.t_order$->{0..1}
          table-strategy:
            inline:
              sharding-column: order_id
              #根据字段值求模
              algorithm-expression: t_order$->{order_id % 2}
          key-generator: #设置主键id生成策略
            column: order_id #主键列
            type: SNOWFLAKE  #设置为雪花算法
            props:
              worker:
                id: 123
              max:
                vibration:
                  offset: 1

 po对象

import com.baomidou.mybatisplus.annotation.*;
import lombok.Data;

import java.io.Serializable;
import java.math.BigDecimal;
@Data
@TableName("t_order") //PS:这里写的是逻辑表的名称
public class OrderPo implements Serializable {
    @TableLogic("t_order")
    @TableId(value = "order_id",type = IdType.AUTO)
    private Long orderId;
    @TableField("order_code")
    private String orderCode;
    @TableField("user_id")
    private Integer userId;
    @TableField("amount")
    private BigDecimal amount;
}

mapper/service不在赘述，正常写法

测试：

    @Test
    void insertOrderDemo(){
        //当前设置的是根据主键除模运算，主键为偶数的落t_order0,奇数的落t_order1
        for (int i = 1; i <= 10; i++) {
            OrderPo orderPo = new OrderPo(preCode + i, new Random().nextInt(10) + i, BigDecimal.TEN);
            orderService.insert(orderPo);
        }
    }

 inline表达式相对比较方便，不需要自己开发额外的分片算法。

列出三种常用的方式

        #根据字段求模分片
        t_order:
          actualDataNodes: ds0.t_order$->{0..1}
          table-strategy:
            inline:
              sharding-column: order_id
              #根据字段值求模
              algorithm-expression: t_order$->{order_id % 2}
              #字段的hashcode求模
              #algorithm-expression: t_order$->{Math.abs(order_id.hashCode()) % 2}


        #根据字段哈希值进行求模分片
        t_order:
          actualDataNodes: ds0.t_order$->{0..1}
          table-strategy:
            inline:
              sharding-column: order_id
              #字段的hashcode求模
              algorithm-expression: t_order$->{Math.abs(order_id.hashCode()) % 2}
        #根据枚举值分片，如订单根据北京、上海等区域或订单类型进行分表时就很适用
        t_area:
          actualDataNodes: ds0.t_area_$->{['shanghai','tianjin']}
          table-strategy:
            inline:
              sharding-column: area
              #枚举算法
              algorithm-expression: t_area_$->{area}