标签：bsp rds hdfs written shuf proc ons snapshot 虚拟
1：首先搞好实体类对象：
write 是把每个对象序列化到输出流，readFields是把输入流字节反序列化，实现WritableComparable，Java值对象的比较：一般需要重写toString(),hashCode(),equals()方法
1 package com.areapartition; 2 3 import java.io.DataInput; 4 import java.io.DataOutput; 5 import java.io.IOException; 6 7 import org.apache.hadoop.io.Writable; 8 import org.apache.hadoop.io.WritableComparable; 9 10 /*** 11 * 12 * @author Administrator 13 * 1:write 是把每个对象序列化到输出流 14 * 2:readFields是把输入流字节反序列化 15 * 3:实现WritableComparable 16 * Java值对象的比较：一般需要重写toString(),hashCode(),equals()方法 17 * 18 */ 19 public class FlowBean implements WritableComparable<FlowBean>{ 20 21 22 private String phoneNumber;//电话号码 23 private long upFlow;//上行流量 24 private long downFlow;//下行流量 25 private long sumFlow;//总流量 26 27 28 29 public String getPhoneNumber() { 30 return phoneNumber; 31 } 32 public void setPhoneNumber(String phoneNumber) { 33 this.phoneNumber = phoneNumber; 34 } 35 public long getUpFlow() { 36 return upFlow; 37 } 38 public void setUpFlow(long upFlow) { 39 this.upFlow = upFlow; 40 } 41 public long getDownFlow() { 42 return downFlow; 43 } 44 public void setDownFlow(long downFlow) { 45 this.downFlow = downFlow; 46 } 47 public long getSumFlow() { 48 return sumFlow; 49 } 50 public void setSumFlow(long sumFlow) { 51 this.sumFlow = sumFlow; 52 } 53 54 //为了对象数据的初始化方便，加入一个带参的构造函数 55 public FlowBean(String phoneNumber, long upFlow, long downFlow) { 56 this.phoneNumber = phoneNumber; 57 this.upFlow = upFlow; 58 this.downFlow = downFlow; 59 this.sumFlow = upFlow + downFlow; 60 } 61 //在反序列化时候，反射机制需要调用空参的构造函数，所以定义了一个空参的构造函数 62 public FlowBean() { 63 } 64 65 //重写toString()方法 66 @Override 67 public String toString() { 68 return "" + upFlow + "\t" + downFlow + "\t" + sumFlow + ""; 69 } 70 71 72 //从数据流中反序列出对象的数据 73 //从数据流中读取字段时必须和序列化的顺序保持一致 74 @Override 75 public void readFields(DataInput in) throws IOException { 76 phoneNumber = in.readUTF(); 77 upFlow = in.readLong(); 78 downFlow = in.readLong(); 79 sumFlow = in.readLong(); 80 81 } 82 83 //将对象数据序列化到流中 84 @Override 85 public void write(DataOutput out) throws IOException { 86 out.writeUTF(phoneNumber); 87 out.writeLong(upFlow); 88 out.writeLong(downFlow); 89 out.writeLong(sumFlow); 90 91 } 92 93 //流量比较的实现方法 94 @Override 95 public int compareTo(FlowBean o) { 96 97 //大就返回-1，小于等于返回1，进行倒序排序 98 return sumFlow > o.sumFlow ? -1 : 1; 99 } 100 101 102 103 }
2：流量分区处理操作的步骤：
2. 1：对流量原始日志进行流量统计，将不同的省份的用户统计结果输出到不同文件；
2.2：需要自定义改造两个机制：
2.2.1：改造分区的逻辑，自定义一个partitioner
2.2.2：自定义reducer task的并发任务数
1 package com.areapartition; 2 3 import java.io.IOException; 4 5 import org.apache.commons.lang.StringUtils; 6 import org.apache.hadoop.conf.Configuration; 7 import org.apache.hadoop.fs.Path; 8 import org.apache.hadoop.io.LongWritable; 9 import org.apache.hadoop.io.Text; 10 import org.apache.hadoop.mapreduce.Job; 11 import org.apache.hadoop.mapreduce.Mapper; 12 import org.apache.hadoop.mapreduce.Reducer; 13 import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; 14 import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat; 15 16 /*** 17 * 流量分区处理操作 18 * @author Administrator 19 * 1:对流量原始日志进行流量统计，将不同的省份的用户统计结果输出到不同文件； 20 * 2:需要自定义改造两个机制： 21 * 2.1:改造分区的逻辑，自定义一个partitioner 22 * 2.2:自定义reducer task的并发任务数 23 */ 24 public class FlowSumArea { 25 26 27 public static class FlowSumAreaMapper extends Mapper<LongWritable, Text, Text, FlowBean>{ 28 @Override 29 protected void map(LongWritable key, Text value, Context context) 30 throws IOException, InterruptedException { 31 //拿到一行数据 32 String line = value.toString(); 33 //切分成各个字段 34 String[] fields = StringUtils.split(line, "\t"); 35 36 //获取到我们需要的字段 37 String phoneNumber = fields[1]; 38 long up_flow = Long.parseLong(fields[7]); 39 long down_flow = Long.parseLong(fields[8]); 40 41 //封装成key-value并且输出 42 context.write(new Text(phoneNumber), new FlowBean(phoneNumber, up_flow, down_flow)); 43 } 44 } 45 46 47 public static class FlowSumAreaReducer extends Reducer<Text, FlowBean, Text, FlowBean>{ 48 @Override 49 protected void reduce(Text key, Iterable<FlowBean> values, Context context) 50 throws IOException, InterruptedException { 51 //遍历求和 52 long up_flowSum = 0; 53 long down_flowSum = 0; 54 for(FlowBean fb : values){ 55 up_flowSum += fb.getUpFlow(); 56 down_flowSum += fb.getDownFlow(); 57 } 58 59 //封装成key-value并且输出 60 context.write(key, new FlowBean(key.toString(),up_flowSum,down_flowSum)); 61 } 62 63 } 64 65 66 public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException { 67 //创建配置文件 68 Configuration conf = new Configuration(); 69 //获取一个作业 70 Job job = Job.getInstance(conf); 71 72 //设置整个job所用的那些类在哪个jar包 73 job.setJarByClass(FlowSumArea.class); 74 //本job使用的mapper和reducer的类 75 job.setMapperClass(FlowSumAreaMapper.class); 76 job.setReducerClass(FlowSumAreaReducer.class); 77 78 //设置我们自定义的分组逻辑定义 79 job.setPartitionerClass(AreaPartitioner.class); 80 81 //指定mapper的输出数据key-value类型 82 job.setMapOutputKeyClass(Text.class); 83 job.setMapOutputValueClass(FlowBean.class); 84 85 //指定reduce的输出数据key-value类型Text 86 job.setOutputKeyClass(Text.class); 87 job.setOutputValueClass(FlowBean.class); 88 89 90 //设置reduce的任务并发数，应该跟分组的数量保持一致 91 job.setNumReduceTasks(7); 92 93 //指定要处理的输入数据存放路径 94 //FileInputFormat是所有以文件作为数据源的InputFormat实现的基类， 95 //FileInputFormat保存作为job输入的所有文件，并实现了对输入文件计算splits的方法。 96 //至于获得记录的方法是有不同的子类——TextInputFormat进行实现的。 97 FileInputFormat.setInputPaths(job, new Path(args[0])); 98 99 //指定处理结果的输出数据存放路径 100 FileOutputFormat.setOutputPath(job, new Path(args[1])); 101 102 //将job提交给集群运行 103 //job.waitForCompletion(true); 104 //正常执行成功返回0，否则返回1 105 System.exit(job.waitForCompletion(true) ? 0 : 1);; 106 107 } 108 109 }
3：从key中拿到手机号，查询手机归属地字典，不同的省份返回不同的组号：
3.1：Partitioner是partitioner的基类，如果需要定制partitioner也需要继承该类。
3.2：HashPartitioner是mapreduce的默认partitioner。计算方法是 which reducer=(key.hashCode() & Integer.MAX_VALUE) % numReduceTasks，得到当前的目的reducer。
1 package com.areapartition; 2 3 import java.util.HashMap; 4 5 import org.apache.hadoop.mapreduce.Partitioner; 6 7 public class AreaPartitioner<KEY,VALUE> extends Partitioner<KEY, VALUE>{ 8 9 private static HashMap<String, Integer> areaMap = new HashMap<String,Integer>(); 10 11 static{ 12 areaMap.put("135", 0); 13 areaMap.put("136", 1); 14 areaMap.put("137", 2); 15 areaMap.put("138", 3); 16 areaMap.put("139", 4); 17 areaMap.put("841", 5); 18 } 19 20 @Override 21 public int getPartition(KEY key, VALUE value, int numPartitions) { 22 //从key中拿到手机号，查询手机归属地字典，不同的省份返回不同的组号 23 Integer areaCoder = areaMap.get(key.toString().subSequence(0, 3)) == null ? 6 : areaMap.get(key.toString().subSequence(0, 3)); 24 25 26 return areaCoder; 27 } 28 29 30 }
4：将打好的jar包上传到虚拟机上面：
然后启动搭建的集群start-dfs.sh，start-yarn.sh：
然后操作如下所示：
1 [root@master hadoop]# hadoop jar flowarea.jar com.areapartition.FlowSumArea /flow/data /flow/areaoutput4 2 17/09/25 15:36:38 INFO client.RMProxy: Connecting to ResourceManager at master/192.168.0.55:8032 3 17/09/25 15:36:38 WARN mapreduce.JobSubmitter: Hadoop command-line option parsing not performed. Implement the Tool interface and execute your application with ToolRunner to remedy this. 4 17/09/25 15:36:38 INFO input.FileInputFormat: Total input paths to process : 1 5 17/09/25 15:36:38 INFO mapreduce.JobSubmitter: number of splits:1 6 17/09/25 15:36:38 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1506324201206_0004 7 17/09/25 15:36:38 INFO impl.YarnClientImpl: Submitted application application_1506324201206_0004 8 17/09/25 15:36:38 INFO mapreduce.Job: The url to track the job: http://master:8088/proxy/application_1506324201206_0004/ 9 17/09/25 15:36:38 INFO mapreduce.Job: Running job: job_1506324201206_0004 10 17/09/25 15:36:43 INFO mapreduce.Job: Job job_1506324201206_0004 running in uber mode : false 11 17/09/25 15:36:43 INFO mapreduce.Job: map 0% reduce 0% 12 17/09/25 15:36:48 INFO mapreduce.Job: map 100% reduce 0% 13 17/09/25 15:36:56 INFO mapreduce.Job: map 100% reduce 14% 14 17/09/25 15:37:04 INFO mapreduce.Job: map 100% reduce 29% 15 17/09/25 15:37:08 INFO mapreduce.Job: map 100% reduce 43% 16 17/09/25 15:37:10 INFO mapreduce.Job: map 100% reduce 71% 17 17/09/25 15:37:11 INFO mapreduce.Job: map 100% reduce 86% 18 17/09/25 15:37:12 INFO mapreduce.Job: map 100% reduce 100% 19 17/09/25 15:37:12 INFO mapreduce.Job: Job job_1506324201206_0004 completed successfully 20 17/09/25 15:37:12 INFO mapreduce.Job: Counters: 49 21 File System Counters 22 FILE: Number of bytes read=1158 23 FILE: Number of bytes written=746635 24 FILE: Number of read operations=0 25 FILE: Number of large read operations=0 26 FILE: Number of write operations=0 27 HDFS: Number of bytes read=2322 28 HDFS: Number of bytes written=526 29 HDFS: Number of read operations=24 30 HDFS: Number of large read operations=0 31 HDFS: Number of write operations=14 32 Job Counters 33 Launched map tasks=1 34 Launched reduce tasks=7 35 Data-local map tasks=1 36 Total time spent by all maps in occupied slots (ms)=2781 37 Total time spent by all reduces in occupied slots (ms)=98540 38 Total time spent by all map tasks (ms)=2781 39 Total time spent by all reduce tasks (ms)=98540 40 Total vcore-seconds taken by all map tasks=2781 41 Total vcore-seconds taken by all reduce tasks=98540 42 Total megabyte-seconds taken by all map tasks=2847744 43 Total megabyte-seconds taken by all reduce tasks=100904960 44 Map-Reduce Framework 45 Map input records=22 46 Map output records=22 47 Map output bytes=1072 48 Map output materialized bytes=1158 49 Input split bytes=93 50 Combine input records=0 51 Combine output records=0 52 Reduce input groups=21 53 Reduce shuffle bytes=1158 54 Reduce input records=22 55 Reduce output records=21 56 Spilled Records=44 57 Shuffled Maps =7 58 Failed Shuffles=0 59 Merged Map outputs=7 60 GC time elapsed (ms)=1751 61 CPU time spent (ms)=4130 62 Physical memory (bytes) snapshot=570224640 63 Virtual memory (bytes) snapshot=2914865152 64 Total committed heap usage (bytes)=234950656 65 Shuffle Errors 66 BAD_ID=0 67 CONNECTION=0 68 IO_ERROR=0 69 WRONG_LENGTH=0 70 WRONG_MAP=0 71 WRONG_REDUCE=0 72 File Input Format Counters 73 Bytes Read=2229 74 File Output Format Counters 75 Bytes Written=526 76 [root@master hadoop]# hadoop fs -ls /flow/ 77 Found 10 items 78 drwxr-xr-x - root supergroup 0 2017-09-25 15:25 /flow/areaoutput 79 drwxr-xr-x - root supergroup 0 2017-09-25 15:34 /flow/areaoutput2 80 drwxr-xr-x - root supergroup 0 2017-09-25 15:35 /flow/areaoutput3 81 drwxr-xr-x - root supergroup 0 2017-09-25 15:37 /flow/areaoutput4 82 -rw-r--r-- 1 root supergroup 2229 2017-09-20 10:00 /flow/data 83 drwxr-xr-x - root supergroup 0 2017-09-20 09:35 /flow/output 84 drwxr-xr-x - root supergroup 0 2017-09-20 09:47 /flow/output2 85 drwxr-xr-x - root supergroup 0 2017-09-20 10:01 /flow/output3 86 drwxr-xr-x - root supergroup 0 2017-09-20 10:21 /flow/output4 87 drwxr-xr-x - root supergroup 0 2017-09-21 19:32 /flow/sortoutput 88 [root@master hadoop]# hadoop fs -ls /flow/areaoutput4 89 Found 8 items 90 -rw-r--r-- 1 root supergroup 0 2017-09-25 15:37 /flow/areaoutput4/_SUCCESS 91 -rw-r--r-- 1 root supergroup 77 2017-09-25 15:36 /flow/areaoutput4/part-r-00000 92 -rw-r--r-- 1 root supergroup 49 2017-09-25 15:37 /flow/areaoutput4/part-r-00001 93 -rw-r--r-- 1 root supergroup 104 2017-09-25 15:37 /flow/areaoutput4/part-r-00002 94 -rw-r--r-- 1 root supergroup 22 2017-09-25 15:37 /flow/areaoutput4/part-r-00003 95 -rw-r--r-- 1 root supergroup 102 2017-09-25 15:37 /flow/areaoutput4/part-r-00004 96 -rw-r--r-- 1 root supergroup 24 2017-09-25 15:37 /flow/areaoutput4/part-r-00005 97 -rw-r--r-- 1 root supergroup 148 2017-09-25 15:37 /flow/areaoutput4/part-r-00006 98 [root@master hadoop]# hadoop fs -cat /flow/areaoutput4/part-r-00000 99 13502468823 102 7335 7437 100 13560436666 954 200 1154 101 13560439658 5892 400 6292 102 [root@master hadoop]# hadoop fs -cat /flow/areaoutput4/part-r-00001 103 13602846565 12 1938 1950 104 13660577991 9 6960 6969 105 [root@master hadoop]# hadoop fs -cat /flow/areaoutput4/part-r-00002 106 13719199419 0 200 200 107 13726230503 2481 24681 27162 108 13726238888 2481 24681 27162 109 13760778710 120 200 320 110 [root@master hadoop]# hadoop fs -cat /flow/areaoutput4/part-r-00003 111 13826544101 0 200 200 112 [root@master hadoop]# hadoop fs -cat /flow/areaoutput4/part-r-00004 113 13922314466 3008 3720 6728 114 13925057413 63 11058 11121 115 13926251106 0 200 200 116 13926435656 1512 200 1712 117 [root@master hadoop]# hadoop fs -cat /flow/areaoutput4/part-r-00005 118 84138413 4116 1432 5548 119 [root@master hadoop]# hadoop fs -cat /flow/areaoutput4/part-r-00006 120 13480253104 180 200 380 121 15013685858 27 3659 3686 122 15920133257 20 3156 3176 123 15989002119 3 1938 1941 124 18211575961 12 1527 1539 125 18320173382 18 9531 9549
5：复制多份测试数据操作如下，测试map的多线程执行：
5.1：map task 的并发数是切片的数量决定的，有多少个切片，就启动多少个map task。
5.2：切片是一个逻辑的概念，指的就是文件中数据的偏移量的范围。
5.3：切片的具体大小应该根据所处理的文件的大小来调整。
[root@master hadoop]# hadoop fs -mkdir /flow/data/ [root@master hadoop]# hadoop fs -put HTTP_20130313143750.dat /flow/data/ [root@master hadoop]# hadoop fs -cp /flow/data/HTTP_20130313143750.dat /flow/data/HTTP_20130313143750.dat.2 [root@master hadoop]# hadoop fs -cp /flow/data/HTTP_20130313143750.dat /flow/data/HTTP_20130313143750.dat.3 [root@master hadoop]# hadoop fs -cp /flow/data/HTTP_20130313143750.dat /flow/data/HTTP_20130313143750.dat.4 [root@master hadoop]# hadoop fs -ls /flow/data/ Found 4 items -rw-r--r-- 1 root supergroup 2229 2017-09-25 16:36 /flow/data/HTTP_20130313143750.dat -rw-r--r-- 1 root supergroup 2229 2017-09-25 16:36 /flow/data/HTTP_20130313143750.dat.2 -rw-r--r-- 1 root supergroup 2229 2017-09-25 16:37 /flow/data/HTTP_20130313143750.dat.3 -rw-r--r-- 1 root supergroup 2229 2017-09-25 16:37 /flow/data/HTTP_20130313143750.dat.4 [root@master hadoop]#
6：Combiners编程
6.1：每一个map可能会产生大量的输出，combiner的作用就是在map端对输出先做一次合并，以减少传输到reducer的数据量。
6.2：combiner最基本是实现本地key的归并，combiner具有类似本地的reduce功能。
6.3： 如果不用combiner，那么，所有的结果都是reduce完成，效率会相对低下。使用combiner，先完成的map会在本地聚合，提升速度。
6.4：注意：Combiner的输出是Reducer的输入，如果Combiner是可插拔的，添加Combiner绝不能改变最终的计算结果。所以Combiner只应该用于那种Reduce的输入key/value与输出key/value类型完全一致，且不影响最终结果的场景。比如累加，最大值等。
7：shuffle机制：
7.1：每个map有一个环形内存缓冲区，用于存储任务的输出。默认大小100MB（io.sort.mb属性），一旦达到阀值0.8（io.sort.spill.percent）,一个后台线程把内容写到(spill)磁盘的指定目录（mapred.local.dir）下的新建的一个溢出写文件。
7.2：写磁盘前，要partition,sort。如果有combiner，combine排序后数据。
7.3：等最后记录写完，合并全部溢出写文件为一个分区且排序的文件。
7.4：Reducer通过Http方式得到输出文件的分区。
7.5：TaskTracker为分区文件运行Reduce任务。复制阶段把Map输出复制到Reducer的内存或磁盘。一个Map任务完成，Reduce就开始复制输出。
7.6：排序阶段合并map输出。然后走Reduce阶段。
一脸懵逼学习Hadoop中的MapReduce程序中自定义分组的实现
标签：bsp rds hdfs written shuf proc ons snapshot 虚拟
原文地址：http://www.cnblogs.com/biehongli/p/7591518.html
