Sqoop1.x的安装配置方法
这篇文章主要介绍“Sqoop1.x的安装配置方法”,在日常操作中,相信很多人在Sqoop1.x的安装配置方法问题上存在疑惑,小编查阅了各式资料,整理出简单好用的操作方法,希望对大家解答”Sqoop1.x的安装配置方法”的疑惑有所帮助!接下来,请跟着小编一起来学习吧!
10年积累的成都网站建设、成都做网站经验,可以快速应对客户对网站的新想法和需求。提供各种问题对应的解决方案。让选择我们的客户得到更好、更有力的网络服务。我虽然不认识你,你也不认识我。但先网站制作后付款的网站建设流程,更有钟山免费网站建设让你可以放心的选择与我们合作。
一、安装hadoop
hadoop:
sqoop2.x:
http://my.oschina.net/u/204498/blog/518941
二、安装sqoop1.x
1.选择对应的版本
[hadoop@hftclclw0001 ~]$ pwd /home/hadoop [hadoop@hftclclw0001 ~]$ wget [hadoop@hftclclw0001 ~]$ tar -zxvf sqoop-1.4.6.bin__hadoop-2.0.4-alpha.tar.gz [hadoop@hftclclw0001 ~]$ cd sqoop-1.4.6.bin__hadoop-2.0.4-alpha/conf [hadoop@hftclclw0001 conf]$ ls -al total 44 drwx------ 2 hadoop root 4096 Nov 25 04:32 . drwx------ 9 hadoop root 4096 Nov 25 04:20 .. -rw------- 1 hadoop root 818 Apr 27 2015 .gitignore -rw------- 1 hadoop root 3895 Apr 27 2015 oraoop-site-template.xml -rw------- 1 hadoop root 1404 Apr 27 2015 sqoop-env-template.cmd -rwx------ 1 hadoop root 1345 Apr 27 2015 sqoop-env-template.sh -rw------- 1 hadoop root 5531 Apr 27 2015 sqoop-site-template.xml -rw------- 1 hadoop root 5531 Apr 27 2015 sqoop-site.xml [hadoop@hftclclw0001 conf]$ cp sqoop-env-template.sh sqoop-env.sh [hadoop@hftclclw0001 conf]$ vim sqoop-env.sh export HADOOP_COMMON_HOME=/home/hadoop/hadoop-2.7.1 #Set path to where hadoop-*-core.jar is available export HADOOP_MAPRED_HOME=/home/hadoop/hadoop-2.7.1/share/hadoop/mapreduce #set the path to where bin/hbase is available => 可以不用,当使用到HBASE时再配置 #export HBASE_HOME=/home/hadoop/hbase-1.0.1.1 #Set the path to where bin/hive is available => 可以不使用,当使用的HIVE时再配置 #export HIVE_HOME=/home/hadoop/apache-hive-1.2.1-bin #Set the path for where zookeper config dir is #export ZOOCFGDIR= export JAVA_HOME=/usr/java/latest => 要安装JDK,之前安装的JRE,使用时会有问题
2.添加对应的jdbc 驱动,我使用的是MySQL
[hadoop@hftclclw0001 lib]$ pwd /home/hadoop/sqoop-1.4.6.bin__hadoop-2.0.4-alpha/lib [hadoop@hftclclw0001 lib]$ ls -al | grep mysql -rw------- 1 hadoop root 848401 Nov 3 06:41 mysql-connector-java-5.1.25-bin.jar
三、Sqoop 1.x 语法
1.安装mysql(配置相应的repo)
[root@hftclclw0001 opt] yum install mysql-server mysql mysql-client
2.启动并测试,并给root用户添加密码
[root@hftclclw0001 opt] service mysqld start [root@hftclclw0001 opt]# netstat -apn|grep 3306 tcp 0 0 0.0.0.0:3306 0.0.0.0:* LISTEN 24540/mysqld [root@hftclclw0001 opt]# mysql -u root -p Enter password: mysql>
3.准备测试数据
我参考的是Apache Sqoop Cookbook 使用的mysql的
https://github.com/jarcec/Apache-Sqoop-Cookbook
使用github上面的mysql文件,创建sqoop用户,创建sqoop数据库,并新增对应的tables。并给sqoop用户赋予相应的权限 mysql> show databases; +--------------------+ | Database | +--------------------+ | information_schema | | mysql | | performance_schema | | sqoop | +--------------------+ 4 rows in set (0.00 sec) mysql> use sqoop mysql> show tables; +-----------------+ | Tables_in_sqoop | +-----------------+ | cities | | countries | | normcities | | staging_cities | | visits | +-----------------+ 5 rows in set (0.00 sec)
chapter2 importing data
sqoop list: [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop-list-tables --connect jdbc:mysql://{ip}:{por}/sqoop \ > --username sqoop \ > --password sqoop ... ... cities countries normcities staging_cities visits => 这些tables就是之前mysql中的新建的
sqoop import:全表导入(transferring an entire table) [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop import \ > --connect jdbc:mysql://{ip}:{port}/sqoop \ > --username sqoop \ > --password sqoop \ > --table cities ... => 会调用MR,读取mysql,并写入文件中(默认理解是当前用户下,table名称对应的木) 总共三条记录,生成了三个文件 [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ hadoop dfs -ls /user/hadoop/cities -rw-r--r-- 3 hadoop supergroup 0 2015-11-25 05:29 /user/hadoop/cities/_SUCCESS -rw-r--r-- 3 hadoop supergroup 16 2015-11-25 05:29 /user/hadoop/cities/part-m-00000 -rw-r--r-- 3 hadoop supergroup 22 2015-11-25 05:29 /user/hadoop/cities/part-m-00001 -rw-r--r-- 3 hadoop supergroup 16 2015-11-25 05:29 /user/hadoop/cities/part-m-00002 [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ hadoop dfs -cat cities/part-m-00000 1,USA,Palo Alto
sqoop import:指定路径(specifying a target directory) --target-dir 指定路径不能存在(针对单表使用的) [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop import \ > --connect jdbc:mysql://{ip}:{port}/sqoop \ > --username sqoop \ > --password sqoop \ > --table cities \ > --target-dir /tmp/cities [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ hadoop dfs -ls /tmp/cities -rw-r--r-- 3 hadoop supergroup 0 2015-11-25 05:29 /user/hadoop/cities/_SUCCESS -rw-r--r-- 3 hadoop supergroup 16 2015-11-25 05:29 /user/hadoop/cities/part-m-00000 -rw-r--r-- 3 hadoop supergroup 22 2015-11-25 05:29 /user/hadoop/cities/part-m-00001 -rw-r--r-- 3 hadoop supergroup 16 2015-11-25 05:29 /user/hadoop/cities/part-m-00002 [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ hadoop dfs -cat cities/part-m-00000 1,USA,Palo Alto 当多表导入是,可以使用--warehouse-dir 会再指定目录下,再生成以table表名称的目录
sqoop import:带where条件的sql,即子集 (importing only a subset of data) mysql> select * from sqoop.cities; +----+----------------+-----------+ | id | country | city | +----+----------------+-----------+ | 1 | USA | Palo Alto | | 2 | Czech Republic | Brno | | 3 | USA | Sunnyvale | +----+----------------+-----------+ 3 rows in set (0.00 sec) mysql> select * from sqoop.cities where country = 'USA'; +----+---------+-----------+ | id | country | city | +----+---------+-----------+ | 1 | USA | Palo Alto | | 3 | USA | Sunnyvale | +----+---------+-----------+ 2 rows in set (0.00 sec) [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ hadoop dfs -rmr /tmp/cities [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop import \ > --connect jdbc:mysql://{ip}:{port}/sqoop \ > --username sqoop \ > --password sqoop \ > --table cities \ > --where "country = 'USA'" \ > --target-dir /tmp/cities
sqoop import:(protecting your password) [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop import \ > --connect jdbc:mysql://{ip}:{port}/sqoop \ > --username sqoop \ > --table cities \ > --where "country = 'USA'" \ > --target-dir /tmp/cities \ > -P => 命令行输入 >--password-file my-sqoop-password => 指定密码文件 sqoop import:(Using a File Format Other Than CSV) 默认生成的是CSV文件,字段间使用tab间隔 --as-sequencefile --as-avrodatafile sqoop import:(Compressing imported data) --compress --compression-codec org.apache.hadoop.io.compress.BZip2Codec =>指定压缩算法
sqoop import:(speeding up transfers) 默认mr的inputformat是通过jdbc的形式读取数据,效率低,可以使用数据库提供的一些工具,如mysql的 mysqldump等 --direct
chapter3 Incremental Import
mysql> select * from sqoop.visits; +----+----------+---------------------+ | id | city | last_update_date | +----+----------+---------------------+ | 1 | Freemont | 1983-05-22 01:01:01 | | 2 | Jicin | 1987-02-02 02:02:02 | +----+----------+---------------------+ 2 rows in set (0.00 sec) importing only new data 表中有个id的主键(int类型的) 我们导入>1的数据 --check-column => 检查那个字段 --last-value => 检查的字段,上次的值是多少,这次会 +1 开始导入 [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop import \ >--connect jdbc:mysql://{ip}:{port}/sqoop \ >--username sqoop \ >--password sqoop \ >--table visits \ >--target-dir /tmp/visits \ >--incremental append \ => incremental 模式是append 即追加 >--check-column id \ => append模式下, 需要一个递增的主键 >--last-value 1 => 会从 id>1开始导入 注意这边在执行的时候是输出以下日志,提示下次增量import是last-value 2 (即本次导入的最后一条记录) 并提示你最好使用 sqoop job --create 来处理类似的定时增量导入 15/11/25 06:05:28 INFO tool.ImportTool: Incremental import complete! To run another incremental import of all data following this import, supply the following arguments: 15/11/25 06:05:28 INFO tool.ImportTool: --incremental append 15/11/25 06:05:28 INFO tool.ImportTool: --check-column id 15/11/25 06:05:28 INFO tool.ImportTool: --last-value 2 15/11/25 06:05:28 INFO tool.ImportTool: (Consider saving this with 'sqoop job --create') [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ hadoop dfs -ls /tmp/visits -rw-r--r-- 3 hadoop supergroup 30 2015-11-25 06:05 /tmp/visits/part-m-00000 [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ hadoop dfs -cat /tmp/visits/part-m-00000 2,Jicin,1987-02-02 02:02:02.0 incrementally importing mutable data
Sqoop Job:
http://shiyanjun.cn/archives/621.html
我们使用Sqoop1.x是,在rdbms和hadoop/hive进行数据同步时,如果是用了--incremental append模式,我们要记录--last-value.如果每次执行同步脚步时候,都需要从日志中解析出一个--last-value的值,然后重新设置脚步的参数,才能正确的保证同步正确。
[hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop job \ >--create visits-sync-job \ => 创建job: job-id(visits-sync-job) >-- \ >import \ >--connect jdbc:mysql://10.224.243.124:3306/sqoop \ >--username sqoop \ >--password sqopp \ >--table visits \ >--incremental append \ >--check-column id \ >--last-value 1 [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop job --list 15/11/25 06:40:00 INFO sqoop.Sqoop: Running Sqoop version: 1.4.6 Available jobs: visits-sync-job [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop job --show visits-sync-job 15/11/25 06:40:10 INFO sqoop.Sqoop: Running Sqoop version: 1.4.6 Enter password: Job: visits-sync-job Tool: import ... ... incremental.last.value = 1 ... [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop job --exec visits-sync-job Enter password: 执行job后,我们在show job [hadoop@hftclclw0001 sqoop-1.4.6.bin__hadoop-2.0.4-alpha]$ ./bin/sqoop job --show visits-sync-job 15/11/25 06:44:52 INFO sqoop.Sqoop: Running Sqoop version: 1.4.6 Enter password: incremental.last.value = 2 => last_value已经被记录下,下次再执行的时候就会读取该值,再执行
chapter 4 Free-Form Query Import
sqoop import:(importing data from two tables) mysql> select * from sqoop.cities; +----+----------------+-----------+ | id | country | city | +----+----------------+-----------+ | 1 | USA | Palo Alto | | 2 | Czech Republic | Brno | | 3 | USA | Sunnyvale | +----+----------------+-----------+ 3 rows in set (0.00 sec) mysql> select * from sqoop.countries; +------------+----------------+ | country_id | country | +------------+----------------+ | 1 | USA | | 2 | Czech Republic | +------------+----------------+ 2 rows in set (0.00 sec)
到此,关于“Sqoop1.x的安装配置方法”的学习就结束了,希望能够解决大家的疑惑。理论与实践的搭配能更好的帮助大家学习,快去试试吧!若想继续学习更多相关知识,请继续关注创新互联网站,小编会继续努力为大家带来更多实用的文章!
本文标题:Sqoop1.x的安装配置方法
分享地址:http://hbruida.cn/article/gdhgse.html