行业知识
Sharding-JDBC之PreciseShardingAlgorithm(精确分片算法)
目录
- 一、简介
- 二、maven依赖
- 三、数据库
- 3.1、创建数据库
- 3.2、创建表
- 四、配置(二选一)
- 4.1、properties配置
- 4.2、yml配置
- 五、精确分片算法
- 5.1、精确分库算法
- 5.2、精确分表算法
- 六、实现
- 6.1、实体层
- 6.2、持久层
- 6.3、服务层
- 6.4、测试类
- 6.4.1、保存订单数据
- 6.4.2、根据订单号查询订单
- 6.4.2、���据订单号和用户查询订单
一、简介
在我之前的文章里,数据的分库分表都是基于行表达式的方式来实现的,看起来也蛮好用,也挺简单的,但是有时会有些复杂的规则,可能使用行表达式策略会很复杂或者实现不了,我们就讲另外一种分片策略,精确分片算法,通常用来处理=或者in条件的情况比较多。
本文示例大概架构如下图:
二、maven依赖
pom.xml
4.0.0 org.springframework.boot spring-boot-starter-parent 2.6.0 com.alian sharding-jdbc 0.0.1-SNAPSHOT sharding-jdbc sharding-jdbc 1.8 org.springframework.boot spring-boot-starter-web org.springframework.boot spring-boot-starter-data-jpa org.apache.shardingsphere sharding-jdbc-spring-boot-starter 4.1.1 com.alibaba druid 1.2.15 mysql mysql-connector-java 8.0.26 runtime org.springframework.boot spring-boot-starter-test test org.apache.commons commons-lang3 3.12.0 org.projectlombok lombok 1.18.20 junit junit 4.12 test org.springframework.boot spring-boot-maven-plugin
有些小伙伴的 druid 可能用的是 druid-spring-boot-starter
com.alibaba druid-spring-boot-starter 1.2.6
然后出现可能使用不了的各种问题,这个时候你只需要在主类上添加 @SpringBootApplication(exclude = {DruidDataSourceAutoConfigure.class}) 即可
package com.alian.shardingjdbc;import com.alibaba.druid.spring.boot.autoconfigure.DruidDataSourceAutoConfigure;import org.springframework.boot.SpringApplication;import org.springframework.boot.autoconfigure.SpringBootApplication;@SpringBootApplication(exclude = {DruidDataSourceAutoConfigure.class})@SpringBootApplicationpublic class ShardingJdbcApplication { public static void main(String[] args) { SpringApplication.run(ShardingJdbcApplication.class, args); }}三、数据库
3.1、创建数据库
CREATE DATABASE `sharding_9` DEFAULT CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci;CREATE DATABASE `sharding_10` DEFAULT CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci;CREATE DATABASE `sharding_11` DEFAULT CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci;
3.2、创建表
在数据库sharding_9、sharding_10、sharding_11下面分别创建两张表:tb_order_1和tb_order_2的结构是一样的
tb_order_1
CREATE TABLE `tb_order_1` ( `order_id` bigint(20) NOT NULL COMMENT '主键', `user_id` int unsigned NOT NULL DEFAULT '0' COMMENT '用户id', `price` int unsigned NOT NULL DEFAULT '0' COMMENT '价格(单位:分)', `order_status` tinyint unsigned NOT NULL DEFAULT '1' COMMENT '订单状态(1:待付款,2:已付款,3:已取消)', `order_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT '创建时间', `title` varchar(100) NOT NULL DEFAULT '' COMMENT '订单标题', PRIMARY KEY (`order_id`), KEY `idx_user_id` (`user_id`), KEY `idx_order_time` (`order_time`)) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COMMENT='订单表';
tb_order_2
CREATE TABLE `tb_order_2` ( `order_id` bigint(20) NOT NULL COMMENT '主键', `user_id` int unsigned NOT NULL DEFAULT '0' COMMENT '用户id', `price` int unsigned NOT NULL DEFAULT '0' COMMENT '价格(单位:分)', `order_status` tinyint unsigned NOT NULL DEFAULT '1' COMMENT '订单状态(1:待付款,2:已付款,3:已取消)', `order_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT '创建时间', `title` varchar(100) NOT NULL DEFAULT '' COMMENT '订单标题', PRIMARY KEY (`order_id`), KEY `idx_user_id` (`user_id`), KEY `idx_order_time` (`order_time`)) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COMMENT='订单表';
四、配置(二选一)
4.1、properties配置
application.properties
server.port=8899server.servlet.context-path=/sharding-jdbc# 允许定义相同的bean对象去覆盖原有的spring.main.allow-bean-definition-overriding=true# 数据源名称,多数据源以逗号分隔spring.shardingsphere.datasource.names=ds1,ds2,ds3# 未配置分片规则的表将通过默认数据源定位spring.shardingsphere.sharding.default-data-source-name=ds1# sharding_9数据库连接池类名称spring.shardingsphere.datasource.ds1.type=com.alibaba.druid.pool.DruidDataSource# sharding_9数据库驱动类名spring.shardingsphere.datasource.ds1.driver-class-name=com.mysql.cj.jdbc.Driver# sharding_9数据库url连接spring.shardingsphere.datasource.ds1.url=jdbc:mysql://192.168.0.129:3306/sharding_9?serverTimezone=GMT%2B8&characterEncoding=utf8&useUnicode=true&useSSL=false&zeroDateTimeBehavior=CONVERT_TO_NULL&autoReconnect=true&allowMultiQueries=true&failOverReadOnly=false&connectTimeout=6000&maxReconnects=5# sharding_9数据库用户名spring.shardingsphere.datasource.ds1.username=alian# sharding_9数据库密码spring.shardingsphere.datasource.ds1.password=123456# sharding_10数据库连接池类名称spring.shardingsphere.datasource.ds2.type=com.alibaba.druid.pool.DruidDataSource# sharding_10数据库驱动类名spring.shardingsphere.datasource.ds2.driver-class-name=com.mysql.cj.jdbc.Driver# sharding_10数据库url连接spring.shardingsphere.datasource.ds2.url=jdbc:mysql://192.168.0.129:3306/sharding_10?serverTimezone=GMT%2B8&characterEncoding=utf8&useUnicode=true&useSSL=false&zeroDateTimeBehavior=CONVERT_TO_NULL&autoReconnect=true&allowMultiQueries=true&failOverReadOnly=false&connectTimeout=6000&maxReconnects=5# sharding_10数据库用户名spring.shardingsphere.datasource.ds2.username=alian# sharding_10数据库密码spring.shardingsphere.datasource.ds2.password=123456# sharding_11数据库连接池类名称spring.shardingsphere.datasource.ds3.type=com.alibaba.druid.pool.DruidDataSource# sharding_11数据库驱动类名spring.shardingsphere.datasource.ds3.driver-class-name=com.mysql.cj.jdbc.Driver# sharding_11数据库url连接spring.shardingsphere.datasource.ds3.url=jdbc:mysql://192.168.0.129:3306/sharding_11?serverTimezone=GMT%2B8&characterEncoding=utf8&useUnicode=true&useSSL=false&zeroDateTimeBehavior=CONVERT_TO_NULL&autoReconnect=true&allowMultiQueries=true&failOverReadOnly=false&connectTimeout=6000&maxReconnects=5# sharding_11数据库用户名spring.shardingsphere.datasource.ds3.username=alian# sharding_11数据库密码spring.shardingsphere.datasource.ds3.password=123456# 采用精确分片策略:PreciseShardingStrategy,根据user_id的奇偶性来添加到不同的库中spring.shardingsphere.sharding.tables.tb_order.database-strategy.standard.sharding-column=user_idspring.shardingsphere.sharding.tables.tb_order.database-strategy.standard.precise-algorithm-class-name=com.alian.shardingjdbc.algorithm.DatabasePreciseShardingAlgorithm# 指定tb_order表的数据分布情况,配置数据节点,使用Groovy的表达式,逻辑表tb_order对应的节点是:ds1.tb_order_1, ds1.tb_order_2,ds2.tb_order_1, ds2.tb_order_2,ds3.tb_order_1, ds3.tb_order_2spring.shardingsphere.sharding.tables.tb_order.actual-data-nodes=ds$->{1..3}.tb_order_$->{1..2}# 采用精确分片策略:PreciseShardingStrategy# 指定tb_order表的分片策略中的分片键spring.shardingsphere.sharding.tables.tb_order.table-strategy.standard.sharding-column=order_id# 指定tb_order表的分片策略中的分片算法表达式,使用Groovy的表达式spring.shardingsphere.sharding.tables.tb_order.table-strategy.standard.precise-algorithm-class-name=com.alian.shardingjdbc.algorithm.OrderTablePreciseShardingAlgorithm# 指定tb_order表的主键为order_idspring.shardingsphere.sharding.tables.tb_order.key-generator.column=order_id# 指定tb_order表的主键生成策略为SNOWFLAKEspring.shardingsphere.sharding.tables.tb_order.key-generator.type=SNOWFLAKE# 指定雪花算法的worker.idspring.shardingsphere.sharding.tables.tb_order.key-generator.props.worker.id=100# 指定雪花算法的max.tolerate.time.difference.millisecondsspring.shardingsphere.sharding.tables.tb_order.key-generator.props.max.tolerate.time.difference.milliseconds=20# 打开sql输出日志spring.shardingsphere.props.sql.show=true4.2、yml配置
application.yml
server: port: 8899 servlet: context-path: /sharding-jdbcspring: main: # 允许定义相同的bean对象去覆盖原有的 allow-bean-definition-overriding: true shardingsphere: props: sql: # 打开sql输出日志 show: true datasource: # 数据源名称,多数据源以逗号分隔 names: ds1,ds2,ds3 ds1: # 数据库连接池类名称 type: com.alibaba.druid.pool.DruidDataSource # 数据库驱动类名 driver-class-name: com.mysql.cj.jdbc.Driver # 数据库url连接 url: jdbc:mysql://192.168.0.129:3306/sharding_9?serverTimezone=GMT%2B8&characterEncoding=utf8&useUnicode=true&useSSL=false&zeroDateTimeBehavior=CONVERT_TO_NULL&autoReconnect=true&allowMultiQueries=true&failOverReadOnly=false&connectTimeout=6000&maxReconnects=5 # 数据库用户名 username: alian # 数据库密码 password: 123456 ds2: # 数据库连接池类名称 type: com.alibaba.druid.pool.DruidDataSource # 数据库驱动类名 driver-class-name: com.mysql.cj.jdbc.Driver # 数据库url连接 url: jdbc:mysql://192.168.0.129:3306/sharding_10?serverTimezone=GMT%2B8&characterEncoding=utf8&useUnicode=true&useSSL=false&zeroDateTimeBehavior=CONVERT_TO_NULL&autoReconnect=true&allowMultiQueries=true&failOverReadOnly=false&connectTimeout=6000&maxReconnects=5 # 数据库用户名 username: alian # 数据库密码 password: 123456 ds3: # 数据库连接池类名称 type: com.alibaba.druid.pool.DruidDataSource # 数据库驱动类名 driver-class-name: com.mysql.cj.jdbc.Driver # 数据库url连接 url: jdbc:mysql://192.168.0.129:3306/sharding_11?serverTimezone=GMT%2B8&characterEncoding=utf8&useUnicode=true&useSSL=false&zeroDateTimeBehavior=CONVERT_TO_NULL&autoReconnect=true&allowMultiQueries=true&failOverReadOnly=false&connectTimeout=6000&maxReconnects=5 # 数据库用户名 username: alian # 数据库密码 password: 123456 sharding: # 未配置分片规则的表将通过默认数据源定位 default-data-source-name: ds1 tables: tb_order: # 由数据源名 + 表名组成,以小数点分隔。多个表以逗号分隔,支持inline表达式 actual-data-nodes: ds$->{1..3}.tb_order_$->{1..2} # 分库策略 database-strategy: # 精确分片策略 standard: # 分片键 sharding-column: user_id # 精确分片算法类名称,用于=和IN precise-algorithm-class-name: com.alian.shardingjdbc.algorithm.DatabasePreciseShardingAlgorithm # 分表策略 table-strategy: # 精确分片策略 standard: # 分片键 sharding-column: order_id # 精确分片算法类名称,用于=和IN precise-algorithm-class-name: com.alian.shardingjdbc.algorithm.OrderTablePreciseShardingAlgorithm # key生成器 key-generator: # 自增列名称,缺省表示不使用自增主键生成器 column: order_id # 自增列值生成器类型,缺省表示使用默认自增列值生成器(SNOWFLAKE/UUID) type: SNOWFLAKE # SnowflakeShardingKeyGenerator props: # SNOWFLAKE算法的worker.id worker: id: 100 # SNOWFLAKE算法的max.tolerate.time.difference.milliseconds max: tolerate: time: difference: milliseconds: 20-
通过精确分片算法完成分库分表
-
database-strategy 采用的是 精确分片策略 ,算法实现类是我们自定义的类 com.alian.shardingjdbc.algorithm.DatabasePreciseShardingAlgorithm
-
table-strategy 采用的是 精确分片策略 ,算法实现类是我们自定义的类 com.alian.shardingjdbc.algorithm.OrderTablePreciseShardingAlgorithm
-
actual-data-nodes 使用Groovy的表达式 ds$->{1…3}.tb_order_$->{1…2},对应的数据源是:ds1、 ds2、 ds3,物理表是:tb_order_1、 tb_order_2,组合起来就有6种方式,这里就不一一列举了
-
key-generator :key生成器,需要指定字段和类型,比如这里如果是SNOWFLAKE,最好也配置下props中的两个属性: worker.id 与 max.tolerate.time.difference.milliseconds 属性
五、精确分片算法
在行表示式分片策略中,基本上只需要配置行表示即可,不需要我们开发java,如果有一些比较特殊的要求,表达式很复杂或者是没办法使用表达式,假设我要求根据 userId 进行分库,要满足:
用户id尾数 要分片到数据库 0,8 ds1 1,3,6,9 ds2 2,4,5,7 ds3 使用行表示就很复杂,我们就可以使用自定义分片算法,这里采用精确分片算法。
5.1、精确分库算法
DatabasePreciseShardingAlgorithm.java
@Slf4jpublic class DatabasePreciseShardingAlgorithm implements PreciseShardingAlgorithm { public DatabasePreciseShardingAlgorithm() { } @Override public String doSharding(Collection dataSourceCollection, PreciseShardingValue preciseShardingValue) { // 获取分片键的值 Integer shardingValue = preciseShardingValue.getValue(); // 获取逻辑 String logicTableName = preciseShardingValue.getLogicTableName(); log.info("分片键的值:{},逻辑表:{}", shardingValue, logicTableName); // 对分片键的值对10取模,得到(0-9),我这里就配置了三个库,实际根据需要修改 // 0,8插入到 ds1 // 1,3,6,9插入到 ds2 // 2,4,5,7插入到 ds3 int index = shardingValue % 10; int sourceTarget; if (ArrayUtils.contains(new int[]{0, 8}, index)) { sourceTarget = 1; } else if (ArrayUtils.contains(new int[]{1, 3, 6, 9}, index)) { sourceTarget = 2; } else { sourceTarget = 3; } // 遍历数据源 for (String databaseSource : dataSourceCollection) { // 判断数据源是否存在 if (databaseSource.endsWith(sourceTarget + "")) { return databaseSource; } } // 不存在则抛出异常 throw new UnsupportedOperationException(); }}实际使用也很简单,我们只需要实现接口 PreciseShardingAlgorithm ,需要注意的是这里的类型 Integer 就是分片键 userId 的类型。然后重写方法 doSharding ,这个方法会有两个参数,第一个就是数据源的集合,第二个是分片对象,我们可以获取到 分片键的值 及其 逻辑表 ,具体见上面代码。
分库时就是需要我们通过自定义的算法计算出需要使用的数据源 databaseSource 。
5.2、精确分表算法
OrderTablePreciseShardingAlgorithm.java
@Slf4jpublic class OrderTablePreciseShardingAlgorithm implements PreciseShardingAlgorithm { public OrderTablePreciseShardingAlgorithm() { } @Override public String doSharding(Collection tableCollection, PreciseShardingValue preciseShardingValue) { // 获取分片键的值 Long shardingValue = preciseShardingValue.getValue(); // 取模分表(取模都是从0到collection.size()) long index = shardingValue % tableCollection.size(); // 判断逻辑表名 String logicTableName = preciseShardingValue.getLogicTableName(); // 物理表名 String PhysicalTableName = logicTableName + "_" + (index + 1); log.info("分片键的值:{},物理表名:{}", shardingValue, PhysicalTableName); // 判断是否存在该表 if (tableCollection.contains(PhysicalTableName)) { return PhysicalTableName; } // 不存在则抛出异常 throw new UnsupportedOperationException(); }}精确分表也是要实现接口 PreciseShardingAlgorithm ,需要注意的是这里的 Long 就是分片键 orderId 的类型。然后重写方法 doSharding ,这个方法会有两个参数,第一个就是物理表的集合,第二个是分片对象,我们可以获取到 分片键的值 及其 逻辑表 ,具体见上面代码。
我们就简单取模分片了,不过我们是通过我们自定义方法去实现的,而不是行表示,因为这样你可以很灵活的设计你们的分片算法,比如你们可以使用基因法等等方式去处理,我这里只是为了演示方便。
六、实现
6.1、实体层
Order.java
@Data@Entity@Table(name = "tb_order")public class Order implements Serializable { @Id @GeneratedValue(strategy = GenerationType.IDENTITY) @Column(name = "order_id") private Long orderId; @Column(name = "user_id") private Integer userId; @Column(name = "price") private Integer price; @Column(name = "order_status") private Integer orderStatus; @Column(name = "title") private String title; @Column(name = "order_time") private Date orderTime;}6.2、持久层
OrderRepository.java
public interface OrderRepository extends PagingAndSortingRepository { /** * 根据订单id查询订单 * @param orderId * @return */ Order findOrderByOrderId(Long orderId); /** * 根据订单id和用户id查询订单 * @param orderId * @param userId * @return */ Order findOrderByOrderIdAndUserId(Long orderId,Integer userId);}6.3、服务层
OrderService.java
@Slf4j@Servicepublic class OrderService { @Autowired private OrderRepository orderRepository; public void saveOrder(Order order) { orderRepository.save(order); } public Order queryOrder(Long orderId) { return orderRepository.findOrderByOrderId(orderId); } public Order findOrderByOrderIdAndUserId(Long orderId, Integer userId) { return orderRepository.findOrderByOrderIdAndUserId(orderId, userId); }}6.4、测试类
OrderTests.java
@Slf4j@RunWith(SpringJUnit4ClassRunner.class)@SpringBootTestpublic class OrderTests { @Autowired private OrderService orderService; @Test public void saveOrder() { for (int i = 0; i6.4.1、保存订单数据
效果图:
从上面的数据来看,满足我们分库分表的要求的,实现都是基于我们自定义的算法实现。
6.4.2、根据订单号查询订单
@Test public void queryOrder() { Long orderId = 875112578379300864L; Order order = orderService.queryOrder(orderId); log.info("查询的结果:{}", order); }20:37:23 575 INFO [main]:分片键的值:875112578379300864,物理表名:tb_order_220:37:23 575 INFO [main]:分片键的值:875112578379300864,物理表名:tb_order_220:37:23 575 INFO [main]:分片键的值:875112578379300864,物理表名:tb_order_220:37:23 595 INFO [main]:Logic SQL: select order0_.order_id as order_id1_0_, order0_.order_status as order_st2_0_, order0_.order_time as order_ti3_0_, order0_.price as price4_0_, order0_.title as title5_0_, order0_.user_id as user_id6_0_ from tb_order order0_ where order0_.order_id=?20:37:23 595 INFO [main]:SQLStatement: SelectStatementContext(super=CommonSQLStatementContext(sqlStatement=org.apache.shardingsphere.sql.parser.sql.statement.dml.SelectStatement@28b68067, tablesContext=org.apache.shardingsphere.sql.parser.binder.segment.table.TablesContext@19540247), tablesContext=org.apache.shardingsphere.sql.parser.binder.segment.table.TablesContext@19540247, projectionsContext=ProjectionsContext(startIndex=7, stopIndex=200, distinctRow=false, projections=[ColumnProjection(owner=order0_, name=order_id, alias=Optional[order_id1_0_]), ColumnProjection(owner=order0_, name=order_status, alias=Optional[order_st2_0_]), ColumnProjection(owner=order0_, name=order_time, alias=Optional[order_ti3_0_]), ColumnProjection(owner=order0_, name=price, alias=Optional[price4_0_]), ColumnProjection(owner=order0_, name=title, alias=Optional[title5_0_]), ColumnProjection(owner=order0_, name=user_id, alias=Optional[user_id6_0_])]), groupByContext=org.apache.shardingsphere.sql.parser.binder.segment.select.groupby.GroupByContext@acb1c9c, orderByContext=org.apache.shardingsphere.sql.parser.binder.segment.select.orderby.OrderByContext@1c681761, paginationContext=org.apache.shardingsphere.sql.parser.binder.segment.select.pagination.PaginationContext@411933, containsSubquery=false)20:37:23 595 INFO [main]:Actual SQL: ds1 ::: select order0_.order_id as order_id1_0_, order0_.order_status as order_st2_0_, order0_.order_time as order_ti3_0_, order0_.price as price4_0_, order0_.title as title5_0_, order0_.user_id as user_id6_0_ from tb_order_2 order0_ where order0_.order_id=? ::: [875112578379300864]20:37:23 595 INFO [main]:Actual SQL: ds2 ::: select order0_.order_id as order_id1_0_, order0_.order_status as order_st2_0_, order0_.order_time as order_ti3_0_, order0_.price as price4_0_, order0_.title as title5_0_, order0_.user_id as user_id6_0_ from tb_order_2 order0_ where order0_.order_id=? ::: [875112578379300864]20:37:23 595 INFO [main]:Actual SQL: ds3 ::: select order0_.order_id as order_id1_0_, order0_.order_status as order_st2_0_, order0_.order_time as order_ti3_0_, order0_.price as price4_0_, order0_.title as title5_0_, order0_.user_id as user_id6_0_ from tb_order_2 order0_ where order0_.order_id=? ::: [875112578379300864]20:37:23 640 INFO [main]:查询的结果:Order(orderId=875112578379300864, userId=1009, price=7811, orderStatus=2, title=, orderTime=2023-06-12 20:24:57.0)
从上面的结果我们可以看到当我们查询order_id为 875112578379300864 的记录时,因为我们之前是按 order_id 取模进行的分表,最终得到的是 tb_order_2 ,但是这里根本不知道是哪个库,所以把 ds1、ds2、ds3 都查了一遍,那有什么方法可以改善么?
6.4.2、根据订单号和用户查询订单
@Test public void findOrderByOrderIdAndUserId() { Long orderId = 875112578379300864L; Integer userId=1009; Order order = orderService.findOrderByOrderIdAndUserId(orderId,userId); log.info("查询的结果:{}", order); }20:41:09 242 INFO [main]:分片键的值:1009,逻辑表:tb_order20:41:09 246 INFO [main]:分片键的值:875112578379300864,物理表名:tb_order_220:41:09 264 INFO [main]:Logic SQL: select order0_.order_id as order_id1_0_, order0_.order_status as order_st2_0_, order0_.order_time as order_ti3_0_, order0_.price as price4_0_, order0_.title as title5_0_, order0_.user_id as user_id6_0_ from tb_order order0_ where order0_.order_id=? and order0_.user_id=?20:41:09 264 INFO [main]:SQLStatement: SelectStatementContext(super=CommonSQLStatementContext(sqlStatement=org.apache.shardingsphere.sql.parser.sql.statement.dml.SelectStatement@58d79479, tablesContext=org.apache.shardingsphere.sql.parser.binder.segment.table.TablesContext@102c24d1), tablesContext=org.apache.shardingsphere.sql.parser.binder.segment.table.TablesContext@102c24d1, projectionsContext=ProjectionsContext(startIndex=7, stopIndex=200, distinctRow=false, projections=[ColumnProjection(owner=order0_, name=order_id, alias=Optional[order_id1_0_]), ColumnProjection(owner=order0_, name=order_status, alias=Optional[order_st2_0_]), ColumnProjection(owner=order0_, name=order_time, alias=Optional[order_ti3_0_]), ColumnProjection(owner=order0_, name=price, alias=Optional[price4_0_]), ColumnProjection(owner=order0_, name=title, alias=Optional[title5_0_]), ColumnProjection(owner=order0_, name=user_id, alias=Optional[user_id6_0_])]), groupByContext=org.apache.shardingsphere.sql.parser.binder.segment.select.groupby.GroupByContext@495f7ca4, orderByContext=org.apache.shardingsphere.sql.parser.binder.segment.select.orderby.OrderByContext@700202fa, paginationContext=org.apache.shardingsphere.sql.parser.binder.segment.select.pagination.PaginationContext@141234df, containsSubquery=false)20:41:09 264 INFO [main]:Actual SQL: ds2 ::: select order0_.order_id as order_id1_0_, order0_.order_status as order_st2_0_, order0_.order_time as order_ti3_0_, order0_.price as price4_0_, order0_.title as title5_0_, order0_.user_id as user_id6_0_ from tb_order_2 order0_ where order0_.order_id=? and order0_.user_id=? ::: [875112578379300864, 1009]20:41:09 318 INFO [main]:查询的结果:Order(orderId=875112578379300864, userId=1009, price=7811, orderStatus=2, title=, orderTime=2023-06-12 20:24:57.0)
从上面的结果我们可以看到当我们查询order_id为 875112578379300864 的记录时,用户id为 1009 的记录时,最终直接查询到 ds2.tb_order_2 ,并没有把所有的库都去查了一遍,因为我们的查询条件里有 userId ,会自动计算到对应的数据源,而按 order_id 取模进行的分表会找到对应的表。所以对于这种一个表多个字段同时分库分表的时候,一定要注意这一点,这样的查询能提高效率。
-
下一篇:异地医保备案怎么登录