Spark RDD数据关联过滤案例 - 简书
http://www.jianshu.com/p/f77cb1c93793
前言
最近在专注Spark开发，记录下自己的工作和学习路程，希望能跟大家互相交流成长QQ:86608625 微信:guofei1990123
背景
detail.txt为用户注册信息全表，filter.txt为需要过滤掉的用户的手机号码
要求detail.txt关联filter.txt并去除detail.txt中包含filter.txt中的用户
原始数据
detail.txt数据张三 13007080388 man 山西 NULL 5 2015-11-05 21:22:41李四 13056677799 woman 四川 地推注册邀请 5 2015-10-11 11:32:19王五 13084470421 man 四川 地推注册邀请 6 2015-08-08 08:11:14filter.txt数据1300708038813056677799
实现思路
两份数据通过手机号主键关联并过滤掉不符合规则数据
具体如下：
1.分别加载两份数据
2.解析数据并拼接成key为手机号的元组
3.用户信息leftOuterJoin过滤表
4.过滤掉不符合规则数据并输出
代码实现
packagesparkimportorg.apache.spark.{SparkConf, SparkContext}/*** Created by郭飞on 2016/5/30.*/objectJoinFilterOne {defmain(args: Array[String]) {//创建SparkConf()并设置App名称及Master地址valconf =newSparkConf().setAppName("JoinFilter").setMaster("local")//创建SparkContext，该对象是提交spark App的入口valsc =newSparkContext(conf)//加载源数据valsourceAll = sc.textFile("C://work//data//test//detail.txt")valsourceFilter = sc.textFile("C://work//data//test//filter.txt")//转换RDD//(13007080388,(张三,13007080388,man,山西,NULL,5,2015-11-05 21:22:41)),// (13056677799,(李四,13056677799,woman,四川,地推注册邀请,5,2015-10-11 11:32:19)),// (13084470421,(王五,13084470421,man,四川,地推注册邀请,6,2015-08-08 08:11:14))valrddAll = sourceAll.map(x=>{valline = x.split("\t")valname = line(0)valphone = line(1)valsex = line(2)valaddr = line(3)valdataType = line(4)valdataCode = line(5)valdataDate = line(6)(phone,(name,phone,sex,addr,dataType,dataCode,dataDate))})//(13007080388,delete), (13056677799,delete)valrddFilter = sourceFilter.map(x=>{valline = x.split("\t")valphone = line(0)(phone,"delete")})//(13056677799,张三,13056677799,woman,四川,地推注册邀请,5,2015-10-11 11:32:19,Some(delete)),// (13084470421,李四,13084470421,man,四川,地推注册邀请,6,2015-08-08 08:11:14,None),// (13007080388,王五,13007080388,man,山西,NULL,5,2015-11-05 21:22:41,Some(delete))rddAll.leftOuterJoin(rddFilter).map(x=>{(x._1,x._2._1._1,x._2._1._2,x._2._1._3,x._2._1._4,x._2._1._5,x._2._1._6,x._2._1._7,x._2.2)}).filter(._9==None).saveAsTextFile("C://work//data//result.txt")//println(result.toBuffer)//过滤后//(13084470421,王宇,13084470421,man,四川,地推注册邀请,6,2015-08-08 08:11:14,None)//停止sc，结束该任务sc.stop()}}
文／MichaelFly（简书作者）原文链接：http://www.jianshu.com/p/f77cb1c93793著作权归作者所有，转载请联系作者获得授权，并标注“简书作者”。
