Rajkumar Singh created HIVE-20275:
-------------------------------------
Summary: hive produces incorrect result when using MIN()/MAX() on
varchar with hive.vectorized.reuse.scratch.columns enabled
Key: HIVE-20275
URL: https://issues.apache.org/jira/browse/HIVE-20275
Project: Hive
Issue Type: Bug
Components: Hive
Affects Versions: 3.1.0
Environment: Hive3.1,java8
Reporter: Rajkumar Singh
Steps to reproduce:
{code}
create table testhive3 (name varchar(8), `time` double);
insert into table testhive3 values
('ABC', 1),
('ABC', 2),
('DEF', 1),
('DEF', 2),
('DEF', 1),
('DEF', 2),
('ABC', 1),
('ABC', 2),
('DEF', 1),
('DEF', 2),
('ABC', 1),
('ABC', 2),
('ABC', 1),
('ABC', 2),
('DEF', 1),
('DEF', 2),
('ABC', 1),
('ABC', 2),
('ABC', 1),
('ABC', 2),
('DEF', 1),
('DEF', 2),
('ABC', 1),
('ABC', 2),
('DEF', 1),
('DEF', 2),
( 'ABC', 1),
( NULL, NULL),
( 'ABC', 1),
( 'ABC', 2),
( 'DEF', 1),
('DEF', 2),
('ABC', 1),
( 'ABC', 2),
('ABC', 1),
( 'ABC', 2),
( 'DEF', 1),
('DEF', 2);
select name, `time` from testhive3 where name = 'ABC' group by name, `time`;
+-------+-------+
| name | time |
+-------+-------+
| ABC | 1.0 |
| ABC | 2.0 |
+-------+-------+
select min(name), `time` from testhive3 where name = 'ABC' group by name,
`time`;
+-------+-------+
| _c0 | time |
+-------+-------+
| NULL | 1.0 |
| NULL | 2.0 |
+-------+-------+
set hive.vectorized.reuse.scratch.columns=false;
select min(name), `time` from testhive3 where name = 'ABC' group by name,
`time`;
+------+-------+
| _c0 | time |
+------+-------+
| ABC | 1.0 |
| ABC | 2.0 |
+------+-------+
{code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
