[HIVE-5023] Hive get wrong result when partition has the same path but different schema or authority

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.12.0
• Assignee: Shuaishuai Nie
• Reporter: Shuaishuai Nie
• Created: 07/Aug/13 22:43
• Updated: 15/Oct/13 23:28
• Resolved: 13/Aug/13 12:13

Description

Hive does not differentiate scheme and authority in file uris which cause wrong result when partition has the same path but different schema or authority. Here is a simple repro

partition file path:
asv://[e-mail]/2013-08-05/00/text1.txt
with content "2013-08-05 00:00:00"
asv://[e-mail]/2013-08-05/00/text2.txt
with content "2013-08-05 00:00:20"

CREATE EXTERNAL TABLE IF NOT EXISTS T1 (t STRING) PARTITIONED BY (ProcessDate STRING, Hour STRING, ClusterName STRING) ROW FORMAT DELIMITED FIELDS TERMINATED by '\t' STORED AS TEXTFILE;
ALTER TABLE T1 DROP IF EXISTS PARTITION(processDate='2013-08-05', Hour='00', clusterName ='CLusterA');
ALTER TABLE T1 ADD IF NOT EXISTS PARTITION(processDate='2013-08-05', Hour='00', clusterName ='ClusterA') LOCATION 'asv://[e-mail]/2013-08-05/00';
ALTER TABLE T1 DROP IF EXISTS PARTITION(processDate='2013-08-05', Hour='00', clusterName ='ClusterB');
ALTER TABLE T1 ADD IF NOT EXISTS PARTITION(processDate='2013-08-05', Hour='00', clusterName ='ClusterB') LOCATION 'asv://[e-mail]/2013-08-05/00';

the expect output of the hive query

SELECT ClusterName, t FROM T1 WHERE ProcessDate=’2013-08-05’ AND Hour=’00’;

should be

ClusterA        2013-08-05 00:00:00
ClusterB        2013-08-05 00:00:20

However it is

ClusterA        2013-08-05 00:00:00
ClusterA        2013-08-05 00:00:20

Issue Links

• is part of: HIVE-4275 Hive does not differentiate scheme and authority in file uris