ApacheSqoop是用来实现结构型数据（如关系数据库）和Hadoop之间进行数据迁移的工具。它充分利用了MapReduce的并行特点以批处理的方式加快数据的传输，同时也借助MapReduce实现了容错。项目地址：HYPERLINK"http://sqoop.apache.org/"\t"http://www.tuicool.com/articles/_blank"http://sqoop.apache.org/目前为止，已经演化出了2个版本：sqoop1和sqoop2。sqoop1的最新版本是1.4.5，sqoop2的最新版本是1.99.3；1.99.3和1.4.5是不兼容的，并且功能尚未开发完成，还不适合在生产环境部署。sqoop支持的数据库：Databaseversion--directsupport?connectstringmatchesHSQLDB1.8.0+NoHYPERLINK"http://jdbchsqldb*/"\t"http://www.tuicool.com/articles/_blank"jdbc:hsqldb:*//MySQL5.0+Yesjdbc:mysql://Oracle10.2.0+NoHYPERLINK"http://jdbcoracle*/"\t"http://www.tuicool.com/articles/_blank"jdbc:oracle:*//PostgreSQL8.3+Yes(importonly)jdbc:postgresql:/guojian@localtest:~/work$sudoapt-getinstallsqoopguojian@localtest:~/work$sqoophelpusage:sqoopCOMMAND[ARGS]Availablecommands:codegenGeneratecodetointeractwithdatabaserecordscreate-hive-tableImportatabledefinitionintoHiveevalEvaluateaSQLstatementanddisplaytheresultsexportExportanHDFSdirectorytoadatabasetablehelpListavailablecommandsimportImportatablefromadatabasetoHDFSimport-all-tablesImporttablesfromadatabasetoHDFSjobWorkwithsavedjobslist-databasesListavailabledatabasesonaserverlist-tablesListavailabletablesinadatabasemergeMergeresultsofincrementalimportsmetastoreRunastandaloneSqoopmetastoreversionDisplayversioninformationSee'sqoophelpCOMMAND'forinformationonaspecificcommand.import是将关系数据库迁移到HDFS上guojian@localtest:~/work$sqoopimport--connectjdbc:mysql://192.168.81.176/hivemeta2db--usernameroot-passwordpasswd--tablesdsguojian@localtest:~/work$hadoopfs-ls/user/guojian/sdsFound5items-rw-r--r--3guojiancug_test02014-09-1116:04/user/guojian/sds/_SUCCESS-rw-r--r--3guojiancug_test4832014-09-1116:03/user/guojian/sds/part-m-00000.snappy-rw-r--r--3guojiancug_test5042014-09-1116:04/user/guojian/sds/part-m-00001.snappy-rw-r--r--3guojiancug_test10012014-09-1116:03/user/guojian/sds/part-m-00002.snappy-rw-r--r--3guojiancug_test9522014-09-1116:03/user/guojian/sds/part-m-00003.snappy可以通过--m设置并行数据，即map的数据