分库分表与自定义分表规则

总体介绍

• Oinone的分库分表方案是基于Sharding-JDBC的整合方案，要先具备一些Sharding-JDBC的知识。[Sharding-JDBC]https://shardingsphere.apache.org/document/current/cn/overview/

• 做分库分表前，大家要有一个明确注意的点就是分表字段（也叫均衡字段）的选择，它是非常重要的，与业务场景非常相关。在明确了分库分表字段以后，甚至在功能上都要做一些妥协。比如分库分表字段在查询管理中做为查询条件是必须带上的，不然效率只会更低。

• 分表字段不允许更新，所以代码里更新策略设置类永不更新，并在设置了在页面修改的时候为readonly

配置分表策略

1. 配置ShardingModel模型走分库分表的数据源pamirsSharding
2. 为pamirsSharding配置数据源以及sharding规则
   a. pamirs.sharding.define用于oinone的数据库表创建用
   b. pamirs.sharding.rule用于分表规则配置
3. 为pamirsSharding配置数据源以及sharding规则
   1）指定模型对应数据源

   pamirs:
   framework:
   system:
     system-ds-key: base
     system-models:
       - base.WorkerNode
   data:
     default-ds-key: pamirs
     ds-map:
       base: base
     modelDsMap:
       "[demo.ShardingModel]": pamirsSharding  #配置模型对应的库

   2）分库分表规则配置

   pamirs: 
   sharding:
   define:
     data-sources:
       ds: pamirs
       pamirsSharding: pamirs #申明pamirsSharding库对应的pamirs数据源
     models:
       "[trigger.PamirsSchedule]":
         tables: 0..13
       "[demo.ShardingModel]":
         tables: 0..7
         table-separator: _
   rule:
     pamirsSharding: #配置pamirsSharding库的分库分表规则
       actual-ds:
         - pamirs  #申明pamirsSharding库对应的pamirs数据源
       sharding-rules:
         # Configure sharding rule ，以下配置跟sharding-jdbc配置一致
         - tables:
             demo_core_sharding_model: #demo_core_sharding_model表规则配置
               actualDataNodes: pamirs.demo_core_sharding_model_${0..7}
               tableStrategy:
                 standard:
                   shardingColumn: user_id
                   shardingAlgorithmName: table_inline
           shardingAlgorithms:
             table_inline:
               type: INLINE
               props:
                 algorithm-expression: demo_core_sharding_model_${(Long.valueOf(user_id) % 8)}
       props:
         sql.show: true

自定义规则

• 默认规则即通用的分库分表策略，如按照数据量、哈希等方式进行分库分表；通常默认规则是可以的。
• 但在一些复杂的业务场景下，使用默认规则可能无法满足需求，需要根据实际情况进行自定义。例如，某些业务可能有特定的数据分布模式或者查询特点，需要定制化的分库分表规则来优化数据访问性能或者满足业务需求。在这种情况下，使用自定义规则可以更好地适应业务的需求。

自定义分表规则示例

示例1：按月份分表（DATE_MONTH ）

package pro.shushi.pamirs.demo.core.sharding;

import cn.hutool.core.date.DateUtil;
import com.google.common.collect.Range;
import org.apache.shardingsphere.sharding.api.sharding.standard.PreciseShardingValue;
import org.apache.shardingsphere.sharding.api.sharding.standard.RangeShardingValue;
import org.apache.shardingsphere.sharding.api.sharding.standard.StandardShardingAlgorithm;
import org.springframework.stereotype.Component;
import pro.shushi.pamirs.meta.annotation.fun.extern.Slf4j;

import java.util.*;

/**
 * @author wangxian
 * @version 1.0
 * @description
 */
@Component
@Slf4j
public class DateMonthShardingAlgorithm implements StandardShardingAlgorithm<Date> {

    private Properties props;

    @Override
    public String doSharding(Collection<String> availableTargetNames, PreciseShardingValue<Date> preciseShardingValue) {
        Date date = preciseShardingValue.getValue();
        String suffix = "_" + (DateUtil.month(date) + 1);
        for (String tableName : availableTargetNames) {
            if (tableName.endsWith(suffix)) {
                return tableName;
            }
        }
        throw new IllegalArgumentException("未找到匹配的数据表");
    }

    @Override
    public Collection<String> doSharding(Collection<String> availableTargetNames, RangeShardingValue<Date> rangeShardingValue) {
        List<String> list = new ArrayList<>();
        log.info(rangeShardingValue.toString());
        Range<Date> valueRange = rangeShardingValue.getValueRange();
        Date lowerDate = valueRange.lowerEndpoint();
        Date upperDate = valueRange.upperEndpoint();
        Integer begin = DateUtil.month(lowerDate) + 1;
        Integer end = DateUtil.month(upperDate) + 1;
        TreeSet<String> suffixList = ShardingUtils.getSuffixListForRange(begin, end);
        for (String tableName : availableTargetNames) {
            if (containTableName(suffixList, tableName)) {
                list.add(tableName);
            }
        }
        return list;
    }

    private boolean containTableName(Set<String> suffixList, String tableName) {
        boolean flag = false;
        for (String s : suffixList) {
            if (tableName.endsWith(s)) {
                flag = true;
                break;
            }
        }
        return flag;
    }

    @Override
    public void init() {

    }

    @Override
    public String getType() {
        return "DATE_MONTH";
    }

    @Override
    public Properties getProps() {
        return this.props;
    }

    @Override
    public void setProps(Properties properties) {
        this.props = props;
    }
}

示例2：按特定字段截取去取模分表

package pro.shushi.pamirs.demo.core.sharding;

import org.apache.shardingsphere.sharding.api.sharding.standard.PreciseShardingValue;
import org.apache.shardingsphere.sharding.api.sharding.standard.RangeShardingValue;
import org.apache.shardingsphere.sharding.api.sharding.standard.StandardShardingAlgorithm;
import org.springframework.stereotype.Component;
import pro.shushi.pamirs.meta.annotation.fun.extern.Slf4j;

import java.util.Collection;
import java.util.Properties;

/**
 * @author wangxian
 * @version 1.0
 * @description
 */
@Component
@Slf4j
public class AppUserCodeShardingAlgorithm implements StandardShardingAlgorithm<String> {

    private Properties props;

    @Override
    public String doSharding(Collection<String> availableTargetNames, PreciseShardingValue<String> preciseShardingValue) {
        String appUserCode = preciseShardingValue.getValue();
        String suffix = "_" + Long.parseLong(appUserCode.substring(1)) % 21;
        for (String tableName : availableTargetNames) {
            if (tableName.endsWith(suffix)) {
                return tableName;
            }
        }
        throw new IllegalArgumentException("未找到匹配的数据表");
    }

    @Override
    public Collection<String> doSharding(final Collection<String> availableTargetNames, final RangeShardingValue<String> shardingValue) {
        return availableTargetNames;
    }

    @Override
    public String getType() {
        return "APP_USER_CODE_TYPE";
    }

    @Override
    public Properties getProps() {
        return this.props;
    }

    @Override
    public void setProps(Properties properties) {
        this.props = props;
    }

    @Override
    public void init() {

    }
}

使用自定义分表策略

1）指定模型对应数据源

pamirs:
  framework:
    system:
      system-ds-key: base
      system-models:
        - base.WorkerNode
    data:
      default-ds-key: pamirs_biz
      ds-map:
        base: base
        demo_core: pamirs
      modelDsMap:
        "[demo.record.MsgRecode]": pamirsSharding

2）分库分表规则配置

pamirs:
  sharding:
    define:
      data-sources:
        ds: pamirs
        pamirsSharding: pamirs
      models:
        "[trigger.PamirsSchedule]":
          tables: 0..13
        "[demo.record.MsgRecode]":
          tables: 0..20
          table-separator: _
    rule:
      pamirsSharding:
        actual-ds:
          - pamirs
        sharding-rules:
          - tables:
              demo_core_record_msg_recode:
                actualDataNodes: pamirs.demo_core_record_msg_recode_${0..20}
                tableStrategy:
                  standard:
                    shardingColumn: app_user_code
                    shardingAlgorithmName: app_user_code_table_algorithm
            shardingAlgorithms:
              app_user_code_table_algorithm:
                type: APP_USER_CODE_TYPE

配置自定义规则SPI

在resources/META-INF/services 配置 org.apache.shardingsphere.sharding.spi.ShardingAlgorithm

pro.shushi.pamirs.demo.core.sharding.AppUserCodeShardingAlgorithm
pro.shushi.pamirs.demo.core.sharding.DateMonthShardingAlgorithm