通过Sqoop来传输mysql/oracle/vertica数据至Hive

1.环境:

CDH

SQOOP2

2.操作步骤

1.准备表与主键

表名:test.test_log

主键:sys_log_id

建表语句

CREATE TABLE hbase_test_test_log_0307
(
ACCT_DATE string,
SYS_SEQ_ID string,
MER_CUST_ID string,
INSERTDT string
);

2.设置环境变量

JAVA_HOME=/usr/java/jdk1.8.0_111/
export JAVA_HOME

3.设置hdfs环境与权限

查看目录下文件

hdfs dfs -ls /tmp

删除已有的目录

hdfs dfs -rm -rf /tmp/hive_test_test_log_0307

授权

hadoop dfs -chmod -R 777 /tmp

4.执行导入脚本

sqoop import
--driver com.vertica.jdbc.Driver
--connect jdbc:vertica://192.168.1.1:8888/dbname?searchpath=test
--username user
--password passwd
--query "select * from test.test_log where acct_date between '20160101' and '20160101'and sys_seq_id >1 and \$CONDITIONS "
--hive-import
--create-hive-table
--hive-table hive_test_test_log_0307
--target-dir "/tmp/hive_test_test_log_0307" -m 1

5.查看导入结果与排错

5.确认数据正确

eg:kill异常任务方法

hadoop job -list
hadoop job -kill job_1478510263374_0069