sqoop的安装：

下载地址：http://apache.dataguru.cn/sqoop/1.4.6/

sqoop-1.4.6.bin__hadoop-2.0.4-alpha.tar.gz

sqoop-1.4.6.tar.gz

我的hadoop版本是2.6，所以我下载的是这两个版本。

sqoop-1.4.6.tar.gz不能直接使用，需要编译。编译后把生成的

sqoop-1.4.6.jar 复制到 sqoop_home/lib下面即可。

复制jdbc connector

mysql-connector-java-5.1.32-bin.jar 或版本更高的mysql jdbc connector.

否则会出现某些bug.

sqoop  list-tables --connect jdbc:mysql://namenode01:3306/employees --username hive --password hive   
mysqlurl=jdbc:mysql://namenode01:3306/employees 

sqoop import 
--connect $mysqlurl 
--username hive 
--password hive 
--table departments 
--target-dir /etl/input/departments 

hdfs dfs -cat /etl/input/departments/* 

insert into departments values('d9999','Evan,Test');
Overriding Type Mapping
--map-column-java c1=Float,c2=String,c3=String ...
Sqoop by default uses four concurrent map tasks to transfer data to Hadoop. 

mysqlurl=jdbc:mysql://namenode01:3306/employees 
sqoop import 
--connect $mysqlurl 
--username hive 
--password hive 
--query 'select d.dept_no,d.dept_name,de.from_date, de.to_date, e.* from employees e join dept_emp de on e.emp_no=de.emp_no join departments d on de.dept_no=d.dept_no WHERE $CONDITIONS' 
--split-by d.dept_no 
--target-dir /etl/input/employees 

--compare the result count from mysql and hdfs file
hdfs dfs -cat /etl/input/employees/* | wc -l
if you want to overwrite the data type, you can specify the column=type
--map-column-hive id=STRING,price=DECIMAL 

mysqlurl=jdbc:mysql://namenode01:3306/employees 
sqoop import 
--connect $mysqlurl 
--username hive 
--password hive 
--query 'select d.dept_no,d.dept_name,de.from_date, de.to_date, e.* from employees e join dept_emp de on e.emp_no=de.emp_no join departments d on de.dept_no=d.dept_no WHERE $CONDITIONS' 
--split-by d.dept_no 
--hive-import 
--hive-table test.employees 
--hive-drop-import-delims 
--null-string '\N' --null-non-string '\N' 
--target-dir /tmp/employees