SQOOP是一款开源的工具,主要用于在Hadoop与传统的数据库间进行数据的传递,下面从SQOOP用户手册上摘录一段描述
Sqoopis a tool designed to transfer data between Hadoop and relational databases.You can use Sqoop to import data from a relational database management system(RDBMS) such as MySQL or Oracle into the Hadoop Distributed File System(HDFS),transform the data in Hadoop MapReduce, and then export the data backinto an RDBMS.
SQOOP是Cloudera公司开源的一款在HDFS以及数据库之间传输数据的软件。内部通过JDBC连接HADOOP以及数据库,因此从理论上来讲,只要是支持JDBC的数据库,SQOOP都可以兼容。并且,SQOOP不仅能把数据以文件的形式导入到HDFS上,还可以直接导入数据到HBASE或者HIVE中。
下面是一些性能测试数据,仅供参考:
表名:tb_keywords
行数:11628209
数据文件大小:1.4G
HDFS –> DB
DB -> HDFS
SQOOP
428s
166s
HDFS<->FILE<->DB
209s
105s
从结果上来看,以FILE作为中转方式性能是要高于SQOOP的。原因如下:
1、 本质上SQOOP使用的是JDBC,效率不会比MYSQL自带的到导入\导出工具效率高
2、 以导入数据到DB为例,SQOOP的设计思想是分阶段提交,也就是说假设一个表有1K行,那么它会先读出100行(默认值),然后插入,提交,再读取100行……如此往复
即便如此,SQOOP也是有优势的,比如说使用的便利性,任务执行的容错性等。在一些测试环境中如果需要的话可以考虑把它拿来作为一个工具使用。
下面是一些操作记录
[wanghai01@tc-crm-rd01.tc.baidu.com bin]$ sh export.sh
Fri Sep 23 20:15:47 CST 2011
11/09/23 20:15:48 WARN tool.BaseSqoopTool: Setting your password on the command-line is insecure. Consider using -P instead.
11/09/23 20:15:48 INFO tool.CodeGenTool: Beginning code generation
11/09/23 20:15:48 INFO manager.MySQLManager: Executing SQL statement: SELECT t.* FROM `tb_keyword_data_201104` AS t LIMIT 1
11/09/23 20:15:48 INFO manager.MySQLManager: Executing SQL statement: SELECT t.* FROM `tb_keyword_data_201104` AS t LIMIT 1
11/09/23 20:15:48 INFO orm.CompilationManager: HADOOP_HOME is /home/wanghai01/hadoop/hadoop-0.20.2/bin/..
11/09/23 20:15:48 INFO orm.CompilationManager: Found hadoop core jar at: /home/wanghai01/hadoop/hadoop-0.20.2/bin/../hadoop-0.20.2-core.jar
11/09/23 20:15:49 ERROR orm.CompilationManager: Could not rename /tmp/sqoop-wanghai01/compile/eb16aae87a119b93acb3bc6ea74b5e97/tb_keyword_data_201104.java to /home/wanghai01/cloudera/sqoop-1.2.0-CDH3B4/bin/./tb_keyword_data_201104.java
11/09/23 20:15:49 INFO orm.CompilationManager: Writing jar file: /tmp/sqoop-wanghai01/compile/eb16aae87a119b93acb3bc6ea74b5e97/tb_keyword_data_201104.jar
11/09/23 20:15:49 INFO mapreduce.ExportJobBase: Beginning export of tb_keyword_data_201104
11/09/23 20:15:49 INFO manager.MySQLManager: Executing SQL statement: SELECT t.* FROM `tb_keyword_data_201104` AS t LIMIT 1
11/09/23 20:15:49 INFO input.FileInputFormat: Total input paths to process : 1
11/09/23 20:15:49 INFO input.FileInputFormat: Total input paths to process : 1
11/09/23 20:15:4