[GitHub] [doris] Accelarator commented on issue #14101: [Problem] duplicate表做简单的group by时耗时1min51s

Tue, 08 Nov 2022 20:38:53 -0800 


Accelarator commented on issue #14101:
URL: https://github.com/apache/doris/issues/14101#issuecomment-1308202541

   > 开一下并发试试看呢，set parallel_fragment_exec_instance_num = 8
   
   但是在做自关联的时候，吞吐会低很多，所以结果还是会比较久。同样的join SQL在CK上执行3s以内可以出结果，doris需要44s
   代码如下
   ``` SQL
   select
       a.appid, a.uid, b.province
   from
       (
           -- group by后400w数据
           select 
               appid
               ,uid 
           from 
               app_online_duration_detail
           where 
               totaldate between '2022-11-02 00:00:00' and '2022-11-03 00:00:00'
           group by 
               appid
               ,uid
       ) a
   left join
       (
           select 
               appid
               ,uid
               ,max(province) as province
           from 
               app_online_duration_detail
           where 
               totaldate between '2022-11-02 00:00:00' and '2022-11-03 00:00:00'
           group by 
               appid
               ,uid
       ) b on b.uid = a.uid and b.appid = a.appid
   limit 10
   ```
   
   **第一层profile：**
   
![image](https://user-images.githubusercontent.com/18416349/200739685-1be53930-903e-4f5c-8c3e-12b95b0bbf2a.png)
   
   **第二层profile：**
   
![image](https://user-images.githubusercontent.com/18416349/200739786-235cae19-2b5d-471a-8d1a-fd287cdfdb8e.png)
   
   **第三层profile：**
   
![image](https://user-images.githubusercontent.com/18416349/200739912-2577b7f1-a25a-499d-bf72-b74842c28125.png)
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to