philipse created HIVE-23081:
-------------------------------
Summary: if partition have the same data location,quey will ignore
the former one
Key: HIVE-23081
URL: https://issues.apache.org/jira/browse/HIVE-23081
Project: Hive
Issue Type: Bug
Components: Query Planning
Affects Versions: 2.1.1
Reporter: philipse
Attachments: image-2020-03-26-14-45-22-717.png,
image-2020-03-26-14-45-33-112.png, image-2020-03-26-14-45-44-959.png,
image-2020-03-26-14-46-01-562.png, image-2020-03-26-14-48-02-096.png
Hi
I have a weird thing,the bevior shows like this:
1) create table test1(id int) partitioned by (year int);
2) insert overwrite table test1(partition=2020) select 1;
3) check the location of partition 2020,so as '/A/B';
4) alter table test1 add partition(year=2021) location '/A/B'
5)check the data
!image-2020-03-26-14-45-22-717.png!
6)check partitrions
!image-2020-03-26-14-45-33-112.png!
7)check data from different partitions
!image-2020-03-26-14-46-01-562.png!
8)The question is,when we check the count,the result sames wrong :
hive> select year,count(1) as cnt from test1 group by year;
OK
2021 1
Time taken: 9.524 seconds, Fetched: 1 row(s)
hive>
Can anyone help review whether we have fixed this issue in furture?
Thansk in advance!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
