Oracle listagg如何去重

1.去重思路:利用listagg会忽略null值的特点

2.上SQL:

复制代码
with T_TEST as
 (
  select 'TOOL-001' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'JITAI-1' STATION_ID from dual union all
  select 'TOOL-003' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'JITAI-1' STATION_ID from dual union all
  select 'TEST' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'S1' STATION_ID from dual union all
  select 'TEST' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'S2' STATION_ID from dual union all
  select 'TEST2' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'S1' STATION_ID from dual union all
  select 'TEST2' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'S2' STATION_ID from dual union all
  select 'TOOL-001' EQPT_ID, 123456 ENTITY_GROUP_RRN, 'JITAI-1' STATION_ID from dual union all
  select 'TOOL-003' EQPT_ID, 123456 ENTITY_GROUP_RRN, 'JITAI-1' STATION_ID from dual union all
  select 'TEST' EQPT_ID, 123456 ENTITY_GROUP_RRN, 'S1' STATION_ID from dual union all
  select 'TEST' EQPT_ID, 123456 ENTITY_GROUP_RRN, 'S2' STATION_ID from dual union all
  select 'TEST2' EQPT_ID, 123456 ENTITY_GROUP_RRN, 'S1' STATION_ID from dual union all
  select 'TEST2' EQPT_ID, 123456 ENTITY_GROUP_RRN, 'S2' STATION_ID from dual 
)
select t.entity_group_rrn,
       listagg(decode(rne, 1, t.eqpt_id, null), ',') within group(order by t.eqpt_id) eqpt_ids,
       listagg(decode(rns, 1, t.station_id, null), ',') within group(order by t.station_id) station_ids
from (select row_number() over(partition by t.entity_group_rrn,t.eqpt_id order by rownum) rne,
               row_number() over(partition by t.entity_group_rrn,t.station_id order by rownum) rns,
               t.*
          from T_TEST t) t
group by t.entity_group_rrn;
复制代码

查询结果

 

总结
1、通常情况可以先分组去重,再合并即可,除非遇到同一分组中,多列分别有重复值,才需要上面那样处理
2、还有另一种去重办法,使用 wm_concat + distinct ,如下:

复制代码
with T_TEST as
 (select 'TOOL-001' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'JITAI-1' STATION_ID from dual union all
  select 'TOOL-003' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'JITAI-1' STATION_ID from dual union all
  select 'TEST' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'S1' STATION_ID from dual union all
  select 'TEST' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'S2' STATION_ID from dual union all
  select 'TEST2' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'S1' STATION_ID from dual union all
  select 'TEST2' EQPT_ID, 10493721 ENTITY_GROUP_RRN, 'S2' STATION_ID from dual 
)
select wm_concat(distinct t.EQPT_ID) ids,entity_group_rrn, wm_concat(distinct t.station_id) ids2
from T_TEST t  group by t.entity_group_rrn
复制代码

 但是由于官方不推荐使用 wm_concat ,所以尽量不要用此方法

posted @   每天进步多一点  阅读(925)  评论(0编辑  收藏  举报
相关博文:
阅读排行:
· DeepSeek “源神”启动!「GitHub 热点速览」
· 微软正式发布.NET 10 Preview 1:开启下一代开发框架新篇章
· C# 集成 DeepSeek 模型实现 AI 私有化(本地部署与 API 调用教程)
· DeepSeek R1 简明指南:架构、训练、本地部署及硬件要求
· 2 本地部署DeepSeek模型构建本地知识库+联网搜索详细步骤
历史上的今天:
2022-06-29 JQuery_九大选择器
2022-06-29 jQuery 选择同时包含两个class的元素的实现方法
2021-06-29 开窗函数 SUM() OVER()
2018-06-29 C#操作Sqlite快速入门及相关工具收集
2017-06-29 用sp_lock诊断SQL Sever的死锁问题
2016-06-29 修改CKFinder上传路径
点击右上角即可分享
微信分享提示