Szabolcs Vasas created SQOOP-2986:
-------------------------------------
Summary: Add validation check for --hive-import and --incremental
lastmodified
Key: SQOOP-2986
URL: https://issues.apache.org/jira/browse/SQOOP-2986
Project: Sqoop
Issue Type: Bug
Affects Versions: 1.4.6
Reporter: Szabolcs Vasas
Assignee: Szabolcs Vasas
Fix For: 1.4.7
Sqoop import with --hive-import and --incremental lastmodified options is not
supported, however the application is able to run with these parameters but it
produces unexpected results, the output can contain duplicate rows.
Steps to reproduce the issue:
1) Create the necessary table for example in MySQL:
CREATE TABLE "employees" (
"id" int(11) NOT NULL,
"name" varchar(45) DEFAULT NULL,
"salary" varchar(45) DEFAULT NULL,
"change_date" datetime DEFAULT NULL,
PRIMARY KEY ("id")
) ENGINE=MyISAM DEFAULT CHARSET=latin1;
INSERT INTO `Employees` (`id`,`name`,`salary`,`change_date`) VALUES
(1,'employee1',1000,now());
INSERT INTO `Employees` (`id`,`name`,`salary`,`change_date`) VALUES
(2,'employee2','2000',now());
INSERT INTO `Employees` (`id`,`name`,`salary`,`change_date`) VALUES
(3,'employee3','3000',now());
INSERT INTO `Employees` (`id`,`name`,`salary`,`change_date`) VALUES
(4,'employee4','4000',now());
INSERT INTO `Employees` (`id`,`name`,`salary`,`change_date`) VALUES
(5,'employee5','5000',now());
2) Import the table to Hive
sudo -u hdfs sqoop import --connect jdbc:mysql://servername:3306/sqoop
--username sqoop --password sqoop --table Employees --num-mappers 1
--hive-import --hive-table Employees
3) Update some rows in MySQL:
UPDATE Employees SET salary=1010, change_date=now() where id=1;
UPDATE Employees SET salary=2010, change_date=now() where id=2;
4) Execute the incremental import command:
sudo -u hdfs sqoop import --verbose --connect
jdbc:mysql://servername:3306/sqoop --username sqoop --password sqoop --table
Employees --incremental lastmodified --check-column change_date --merge-key id
--num-mappers 1 --hive-import --hive-table Employees --last-value
"last_timestamp"
5) As a result employees with ids 1 and 2 will not be updated but we will see
duplicate rows in the Hive table.
The task is to introduce a fail-fast validation which will make the Sqoop
import fail if it was submitted with --hive-import and --incremental
lastmodified options.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
