欢迎您访问程序员文章站本站旨在为大家提供分享程序员计算机编程知识!
您现在的位置是: 首页  >  科技

Hadoop实践|矩阵乘法

程序员文章站 2022-09-30 22:08:36
这篇博文列举了两种基础的方法,区别在于输入数据的形式, 一种是矩阵形式;一种是(横坐标,纵坐标,值)的形式 同时这篇博文提到关于矩阵运算的应用: 两种实现的reduce阶段,计算最后结果...

这篇博文列举了两种基础的方法,区别在于输入数据的形式,
一种是矩阵形式;一种是(横坐标,纵坐标,值)的形式

同时这篇博文提到关于矩阵运算的应用:

两种实现的reduce阶段,计算最后结果时,都是直接使用内存存储数据、计算结果,所以当数据量很大的时候(通常都会很大,否则不会用分布式处理),极易造成内存溢出,所以,对于大矩阵的运算,还需要其他的转换方式,比如行列相乘运算、分块矩阵运算、基于最小粒度相乘的算法等方式。另外,因为这两份代码都是demo,所以代码中缺少过滤错误数据的部分。

本文使用的第二种

自己建了两个数组,大小为2" role="presentation">3、 3" role="presentation">3
Hadoop实践|矩阵乘法

你需要做的就是把 输入文件放进去
配置文件复制到工程src文件件下
复制粘贴

package org.apache.hadoop.examples;

import java.io.IOException;  
import java.util.HashMap;  
import java.util.Iterator;  
import java.util.Map;  

import org.apache.hadoop.conf.Configuration;  
import org.apache.hadoop.fs.Path;  
import org.apache.hadoop.io.IntWritable;  
import org.apache.hadoop.io.LongWritable;  
import org.apache.hadoop.io.Text;  
import org.apache.hadoop.mapreduce.Job;  
import org.apache.hadoop.mapreduce.Mapper;  
import org.apache.hadoop.mapreduce.Reducer;  
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;  
import org.apache.hadoop.mapreduce.lib.input.FileSplit;  
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;  
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;  
import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;  

/** 
 * @author liuxinghao 
 * @version 1.0 Created on 2014年10月10日 
 */  
public class SparseMatrixMultiply {  
    public static class SMMapper extends Mapper {  
        private String flag = null;  
        private int m = 2;// 矩阵A的行数  
        private int p = 3;// 矩阵B的列数  

        @Override  
        protected void setup(Context context) throws IOException,  
                InterruptedException {  
            FileSplit split = (FileSplit) context.getInputSplit();  
            flag = split.getPath().getName();  
        }  

        @Override  
        protected void map(LongWritable key, Text value, Context context)  
                throws IOException, InterruptedException {  
            String[] val = value.toString().split(",");  
            if ("t1".equals(flag)) {  
                for (int i = 1; i <= p; i++) {  
                    context.write(new Text(val[0] + "," + i), new Text("a,"  
                            + val[1] + "," + val[2]));  
                }  
            } else if ("t2".equals(flag)) {  
                for (int i = 1; i <= m; i++) {  
                    context.write(new Text(i + "," + val[1]), new Text("b,"  
                            + val[0] + "," + val[2]));  
                }  
            }  
        }  
    }  

    public static class SMReducer extends  
            Reducer {  
        @Override  
        protected void reduce(Text key, Iterable values, Context context)  
                throws IOException, InterruptedException {  
            Map mapA = new HashMap();  
            Map mapB = new HashMap();  

            for (Text value : values) {  
                String[] val = value.toString().split(",");  
                if ("a".equals(val[0])) {  
                    mapA.put(val[1], val[2]);  
                } else if ("b".equals(val[0])) {  
                    mapB.put(val[1], val[2]);  
                }  
            }  

            int result = 0;  
            // 可能在mapA中存在在mapB中不存在的key,或相反情况  
            // 因为,数据定义的时候使用的是稀疏矩阵的定义  
            // 所以,这种只存在于一个map中的key,说明其对应元素为0,不影响结果  
            Iterator mKeys = mapA.keySet().iterator();  
            while (mKeys.hasNext()) {  
                String mkey = mKeys.next();  
                if (mapB.get(mkey) == null) {// 因为mkey取的是mapA的key集合,所以只需要判断mapB是否存在即可。  
                    continue;  
                }  
                result += Integer.parseInt(mapA.get(mkey))  
                        * Integer.parseInt(mapB.get(mkey));  
            }  
            context.write(key, new IntWritable(result));  
        }  
    }  

    public static void main(String[] args) throws IOException,  
            ClassNotFoundException, InterruptedException {  
        String input1 = "hdfs:/matrix/t1";  
        String input2 = "hdfs:/matrix/t2";  
        String output = "hdfs:/matrix/out";  

        Configuration conf = new Configuration();  
        conf.addResource("classpath:/core-site.xml");  
        conf.addResource("classpath:/hdfs-site.xml");  

        Job job = Job.getInstance(conf, "SparseMatrixMultiply");  
        job.setJarByClass(SparseMatrixMultiply.class);  
        job.setOutputKeyClass(Text.class);  
        job.setOutputValueClass(Text.class);  

        job.setMapperClass(SMMapper.class);  
        job.setReducerClass(SMReducer.class);  

        job.setInputFormatClass(TextInputFormat.class);  
        job.setOutputFormatClass(TextOutputFormat.class);  

        FileInputFormat.setInputPaths(job, new Path(input1), new Path(input2));// 加载2个输入数据集  
        Path outputPath = new Path(output);  
        outputPath.getFileSystem(conf).delete(outputPath, true);  
        FileOutputFormat.setOutputPath(job, outputPath);  

        System.exit(job.waitForCompletion(true)  0 : 1);  
    }  
}  

日志文件:

18/03/25 04:24:15 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
18/03/25 04:24:17 INFO Configuration.deprecation: session.id is deprecated. Instead, use dfs.metrics.session-id
18/03/25 04:24:17 INFO jvm.JvmMetrics: Initializing JVM Metrics with processName=JobTracker, sessionId=
18/03/25 04:24:17 WARN mapreduce.JobResourceUploader: Hadoop command-line option parsing not performed. Implement the Tool interface and execute your application with ToolRunner to remedy this.
18/03/25 04:24:17 WARN mapreduce.JobResourceUploader: No job jar file set.  User classes may not be found. See Job or Job#setJar(String).
18/03/25 04:24:17 INFO input.FileInputFormat: Total input paths to process : 2
18/03/25 04:24:17 INFO mapreduce.JobSubmitter: number of splits:2
18/03/25 04:24:18 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_local1549038765_0001
18/03/25 04:24:18 INFO mapreduce.Job: The url to track the job: https://localhost:8080/
18/03/25 04:24:18 INFO mapreduce.Job: Running job: job_local1549038765_0001
18/03/25 04:24:18 INFO mapred.LocalJobRunner: OutputCommitter set in config null
18/03/25 04:24:18 INFO output.FileOutputCommitter: File Output Committer Algorithm version is 1
18/03/25 04:24:18 INFO mapred.LocalJobRunner: OutputCommitter is org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter
18/03/25 04:24:18 INFO mapred.LocalJobRunner: Waiting for map tasks
18/03/25 04:24:18 INFO mapred.LocalJobRunner: Starting task: attempt_local1549038765_0001_m_000000_0
18/03/25 04:24:18 INFO output.FileOutputCommitter: File Output Committer Algorithm version is 1
18/03/25 04:24:18 INFO mapred.Task:  Using ResourceCalculatorProcessTree : [ ]
18/03/25 04:24:18 INFO mapred.MapTask: Processing split: hdfs://master:9000/matrix/t2:0+54
18/03/25 04:24:18 INFO mapred.MapTask: (EQUATOR) 0 kvi 26214396(104857584)
18/03/25 04:24:18 INFO mapred.MapTask: mapreduce.task.io.sort.mb: 100
18/03/25 04:24:18 INFO mapred.MapTask: soft limit at 83886080
18/03/25 04:24:18 INFO mapred.MapTask: bufstart = 0; bufvoid = 104857600
18/03/25 04:24:18 INFO mapred.MapTask: kvstart = 26214396; length = 6553600
18/03/25 04:24:18 INFO mapred.MapTask: Map output collector class = org.apache.hadoop.mapred.MapTask$MapOutputBuffer
18/03/25 04:24:19 INFO mapred.LocalJobRunner: 
18/03/25 04:24:19 INFO mapred.MapTask: Starting flush of map output
18/03/25 04:24:19 INFO mapred.MapTask: Spilling map output
18/03/25 04:24:19 INFO mapred.MapTask: bufstart = 0; bufend = 180; bufvoid = 104857600
18/03/25 04:24:19 INFO mapred.MapTask: kvstart = 26214396(104857584); kvend = 26214328(104857312); length = 69/6553600
18/03/25 04:24:19 INFO mapred.MapTask: Finished spill 0
18/03/25 04:24:19 INFO mapred.Task: Task:attempt_local1549038765_0001_m_000000_0 is done. And is in the process of committing
18/03/25 04:24:19 INFO mapred.LocalJobRunner: map
18/03/25 04:24:19 INFO mapred.Task: Task 'attempt_local1549038765_0001_m_000000_0' done.
18/03/25 04:24:19 INFO mapred.LocalJobRunner: Finishing task: attempt_local1549038765_0001_m_000000_0
18/03/25 04:24:19 INFO mapred.LocalJobRunner: Starting task: attempt_local1549038765_0001_m_000001_0
18/03/25 04:24:19 INFO output.FileOutputCommitter: File Output Committer Algorithm version is 1
18/03/25 04:24:19 INFO mapred.Task:  Using ResourceCalculatorProcessTree : [ ]
18/03/25 04:24:19 INFO mapreduce.Job: Job job_local1549038765_0001 running in uber mode : false
18/03/25 04:24:19 INFO mapred.MapTask: Processing split: hdfs://master:9000/matrix/t1:0+36
18/03/25 04:24:19 INFO mapreduce.Job:  map 100% reduce 0%
18/03/25 04:24:19 INFO mapred.MapTask: (EQUATOR) 0 kvi 26214396(104857584)
18/03/25 04:24:19 INFO mapred.MapTask: mapreduce.task.io.sort.mb: 100
18/03/25 04:24:19 INFO mapred.MapTask: soft limit at 83886080
18/03/25 04:24:19 INFO mapred.MapTask: bufstart = 0; bufvoid = 104857600
18/03/25 04:24:19 INFO mapred.MapTask: kvstart = 26214396; length = 6553600
18/03/25 04:24:19 INFO mapred.MapTask: Map output collector class = org.apache.hadoop.mapred.MapTask$MapOutputBuffer
18/03/25 04:24:19 INFO mapred.LocalJobRunner: 
18/03/25 04:24:19 INFO mapred.MapTask: Starting flush of map output
18/03/25 04:24:19 INFO mapred.MapTask: Spilling map output
18/03/25 04:24:19 INFO mapred.MapTask: bufstart = 0; bufend = 180; bufvoid = 104857600
18/03/25 04:24:19 INFO mapred.MapTask: kvstart = 26214396(104857584); kvend = 26214328(104857312); length = 69/6553600
18/03/25 04:24:19 INFO mapred.MapTask: Finished spill 0
18/03/25 04:24:19 INFO mapred.Task: Task:attempt_local1549038765_0001_m_000001_0 is done. And is in the process of committing
18/03/25 04:24:19 INFO mapred.LocalJobRunner: map
18/03/25 04:24:19 INFO mapred.Task: Task 'attempt_local1549038765_0001_m_000001_0' done.
18/03/25 04:24:19 INFO mapred.LocalJobRunner: Finishing task: attempt_local1549038765_0001_m_000001_0
18/03/25 04:24:19 INFO mapred.LocalJobRunner: map task executor complete.
18/03/25 04:24:19 INFO mapred.LocalJobRunner: Waiting for reduce tasks
18/03/25 04:24:19 INFO mapred.LocalJobRunner: Starting task: attempt_local1549038765_0001_r_000000_0
18/03/25 04:24:19 INFO output.FileOutputCommitter: File Output Committer Algorithm version is 1
18/03/25 04:24:19 INFO mapred.Task:  Using ResourceCalculatorProcessTree : [ ]
18/03/25 04:24:19 INFO mapred.ReduceTask: Using ShuffleConsumerPlugin: org.apache.hadoop.mapreduce.task.reduce.Shuffle@1bc75a85
18/03/25 04:24:20 INFO reduce.MergeManagerImpl: MergerManager: memoryLimit=322594400, maxSingleShuffleLimit=80648600, mergeThreshold=212912320, ioSortFactor=10, memToMemMergeOutputsThreshold=10
18/03/25 04:24:20 INFO reduce.EventFetcher: attempt_local1549038765_0001_r_000000_0 Thread started: EventFetcher for fetching Map Completion Events
18/03/25 04:24:20 INFO reduce.LocalFetcher: localfetcher#1 about to shuffle output of map attempt_local1549038765_0001_m_000001_0 decomp: 218 len: 222 to MEMORY
18/03/25 04:24:20 INFO reduce.InMemoryMapOutput: Read 218 bytes from map-output for attempt_local1549038765_0001_m_000001_0
18/03/25 04:24:20 INFO reduce.MergeManagerImpl: closeInMemoryFile -> map-output of size: 218, inMemoryMapOutputs.size() -> 1, commitMemory -> 0, usedMemory ->218
18/03/25 04:24:20 INFO reduce.LocalFetcher: localfetcher#1 about to shuffle output of map attempt_local1549038765_0001_m_000000_0 decomp: 218 len: 222 to MEMORY
18/03/25 04:24:20 INFO reduce.InMemoryMapOutput: Read 218 bytes from map-output for attempt_local1549038765_0001_m_000000_0
18/03/25 04:24:20 INFO reduce.MergeManagerImpl: closeInMemoryFile -> map-output of size: 218, inMemoryMapOutputs.size() -> 2, commitMemory -> 218, usedMemory ->436
18/03/25 04:24:20 INFO reduce.EventFetcher: EventFetcher is interrupted.. Returning
18/03/25 04:24:20 INFO mapred.LocalJobRunner: 2 / 2 copied.
18/03/25 04:24:20 INFO reduce.MergeManagerImpl: finalMerge called with 2 in-memory map-outputs and 0 on-disk map-outputs
18/03/25 04:24:20 INFO mapred.Merger: Merging 2 sorted segments
18/03/25 04:24:20 INFO mapred.Merger: Down to the last merge-pass, with 2 segments left of total size: 424 bytes
18/03/25 04:24:20 INFO reduce.MergeManagerImpl: Merged 2 segments, 436 bytes to disk to satisfy reduce memory limit
18/03/25 04:24:20 INFO reduce.MergeManagerImpl: Merging 1 files, 438 bytes from disk
18/03/25 04:24:20 INFO reduce.MergeManagerImpl: Merging 0 segments, 0 bytes from memory into reduce
18/03/25 04:24:20 INFO mapred.Merger: Merging 1 sorted segments
18/03/25 04:24:20 INFO mapred.Merger: Down to the last merge-pass, with 1 segments left of total size: 428 bytes
18/03/25 04:24:20 INFO mapred.LocalJobRunner: 2 / 2 copied.
18/03/25 04:24:20 INFO Configuration.deprecation: mapred.skip.on is deprecated. Instead, use mapreduce.job.skiprecords
18/03/25 04:24:20 INFO mapred.Task: Task:attempt_local1549038765_0001_r_000000_0 is done. And is in the process of committing
18/03/25 04:24:20 INFO mapred.LocalJobRunner: 2 / 2 copied.
18/03/25 04:24:20 INFO mapred.Task: Task attempt_local1549038765_0001_r_000000_0 is allowed to commit now
18/03/25 04:24:20 INFO output.FileOutputCommitter: Saved output of task 'attempt_local1549038765_0001_r_000000_0' to hdfs://master:9000/matrix/out/_temporary/0/task_local1549038765_0001_r_000000
18/03/25 04:24:20 INFO mapred.LocalJobRunner: reduce > reduce
18/03/25 04:24:20 INFO mapred.Task: Task 'attempt_local1549038765_0001_r_000000_0' done.
18/03/25 04:24:20 INFO mapred.LocalJobRunner: Finishing task: attempt_local1549038765_0001_r_000000_0
18/03/25 04:24:20 INFO mapred.LocalJobRunner: reduce task executor complete.
18/03/25 04:24:21 INFO mapreduce.Job:  map 100% reduce 100%
18/03/25 04:24:21 INFO mapreduce.Job: Job job_local1549038765_0001 completed successfully
18/03/25 04:24:21 INFO mapreduce.Job: Counters: 35
    File System Counters
        FILE: Number of bytes read=2139
        FILE: Number of bytes written=875557
        FILE: Number of read operations=0
        FILE: Number of large read operations=0
        FILE: Number of write operations=0
        HDFS: Number of bytes read=234
        HDFS: Number of bytes written=42
        HDFS: Number of read operations=28
        HDFS: Number of large read operations=0
        HDFS: Number of write operations=8
    Map-Reduce Framework
        Map input records=15
        Map output records=36
        Map output bytes=360
        Map output materialized bytes=444
        Input split bytes=186
        Combine input records=0
        Combine output records=0
        Reduce input groups=6
        Reduce shuffle bytes=444
        Reduce input records=36
        Reduce output records=6
        Spilled Records=72
        Shuffled Maps =2
        Failed Shuffles=0
        Merged Map outputs=2
        GC time elapsed (ms)=19
        Total committed heap usage (bytes)=907542528
    Shuffle Errors
        BAD_ID=0
        CONNECTION=0
        IO_ERROR=0
        WRONG_LENGTH=0
        WRONG_MAP=0
        WRONG_REDUCE=0
    File Input Format Counters 
        Bytes Read=90
    File Output Format Counters 
        Bytes Written=42