[jira] [Created] (HIVE-16458) Inserting data into hive S3 external table using INSERT INTO SELECT overwrites file

["raj nelur (JIRA)" <ji...@apache.org>]

raj nelur created HIVE-16458:
--------------------------------

             Summary: Inserting data into hive S3 external table using INSERT INTO SELECT overwrites file
                 Key: HIVE-16458
                 URL: https://issues.apache.org/jira/browse/HIVE-16458
             Project: Hive
          Issue Type: Bug
          Components: Hive
    Affects Versions: 1.2.1
         Environment: Hadoop 2.7.3.2.5.3.0-37
Hive 1.2.1000.2.5.3.0-37
OS: RHEL 6.7
            Reporter: raj nelur


I am trying to insert data into the external table using another table (tbl_test) which is not an external table.

insert into tbl_external select * from tbl_test;
But Hive is always generating a file with same file name (000000_0) when I run the above insert statement. As a result, the file is getting overwritten and the S3 location will always have a single file instead of multiple files when I run the insert statements multiple times.

How can I make Hive generate different file names? (eg 000000_0_copy_1, 000000_0_copy_2, etc)

We don't have this problem in our Dev environment, we are facing the problem only in Prod.

Dev Hive version is 1.2.1.2.3.6.0-3796
Prod Hive version is 1.2.1000.2.5.3.0-37



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)