PageRank在Hadoop和spark下的实现以及对比

  1 package org.apache.hadoop.PageRank;
  2 
  3 import java.util.ArrayList;
  4 
  5 import org.apache.hadoop.conf.Configuration;
  6 import org.apache.hadoop.fs.FileSystem;
  7 import org.apache.hadoop.fs.Path;
  8 import org.apache.hadoop.io.Text;
  9 import org.apache.hadoop.mapreduce.Job;
 10 import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
 11 import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
 12 
 13 public class PageRank {
 14 
 15     public static void run(){
 16         
 17     }
 18     
 19     public static void main(String[] args) throws Exception {
 20         double factor=0;
 21         if(args.length>1){
 22             factor=Double.parseDouble(args[0]);
 23         }else{
 24             factor=0.85;
 25         }
 26         String input="hdfs://10.107.8.110:9000/PageRank_input";
 27         String output="hdfs://10.107.8.110:9000/PageRank/output";
 28         ArrayList<String> pathList=new ArrayList<String>();        
 29         for(int i=0;i<20;i++){
 30             Configuration conf = new Configuration();
 31             conf.set("num","4");
 32             conf.set("factor",String.valueOf(factor));
 33             Job job = Job.getInstance(conf, "PageRank");
 34             job.setJarByClass(org.apache.hadoop.PageRank.PageRank.class);
 35             job.setMapperClass(MyMapper.class);
 36             job.setReducerClass(MyReducer.class);
 37             job.setOutputKeyClass(Text.class);
 38             job.setOutputValueClass(Text.class);
 39             FileInputFormat.setInputPaths(job, new Path(input));
 40             FileOutputFormat.setOutputPath(job, new Path(output));
 41             input=output;
 42             pathList.add(output);
 43             output=output+1;
 44             
 45             System.out.println("the "+i+"th iterator is finished");
 46             job.waitForCompletion(true);
 47         }
 48         for(int i=0;i<pathList.size()-1;i++){
 49             Configuration conf=new Configuration();
 50             Path path=new Path(pathList.get(i));
 51             FileSystem fs=path.getFileSystem(conf);
 52             fs.delete(path,true);
 53         }
 54     }
 55 
 56 }
 57 
 58 
 59 
 60 package org.apache.hadoop.PageRank;
 61 
 62 import java.io.IOException;
 63 import java.util.HashMap;
 64 import java.util.Map;
 65 
 66 
 67 import org.apache.hadoop.io.LongWritable;
 68 import org.apache.hadoop.io.Text;
 69 import org.apache.hadoop.mapreduce.Mapper;
 70 
 71 public class MyMapper extends Mapper<LongWritable, Text, Text, Text> {
 72 
 73     
 74     public void map(LongWritable ikey, Text ivalue, Context context)
 75             throws IOException, InterruptedException {
 76         String[] line=ivalue.toString().split(":");
 77         String content=line[1];
 78         int num=content.split(",").length;
 79         String word=line[0].split("    ")[0];
 80         String alaph=line[0].split("    ")[1];
 81         context.write(new Text(word),new Text("content:"+content));
 82         for(String w:content.split(",")){
 83             context.write(new Text(w),new Text("link:"+word+" "+String.valueOf(1.0/num)));
 84             context.write(new Text(w),new Text("alaph:"+word+" "+alaph));
 85         }
 86     }
 87 
 88 }
 89 
 90 
 91 
 92 package org.apache.hadoop.PageRank;
 93 
 94 import java.io.IOException;
 95 import java.util.HashMap;
 96 import java.util.Map;
 97 
 98 import org.apache.hadoop.conf.Configuration;
 99 import org.apache.hadoop.io.Text;
100 import org.apache.hadoop.mapreduce.Reducer;
101 
102 public class MyReducer extends Reducer<Text, Text, Text, Text> {
103 
104     public void reduce(Text _key, Iterable<Text> values, Context context)
105             throws IOException, InterruptedException {
106         // process values
107         Configuration conf=context.getConfiguration();
108         double factor=Double.parseDouble(conf.get("factor"));
109         int num=Integer.parseInt(conf.get("num"));
110         
111         Map<String,Double> alaph=new HashMap<String,Double>();
112         Map<String,Double> link=new HashMap<String,Double>();
113         
114         String content="";
115         for (Text val : values) {
116             String[] line=val.toString().split(":");
117             if(line[0].compareTo("content")==0){
118                 content=line[1];
119             }else {
120                 String[] s=line[1].split(" ");
121                 double d=Double.parseDouble(s[1]);
122                 if(line[0].compareTo("alaph")==0){
123                     alaph.put(s[0],d);
124                 }else if(line[0].compareTo("link")==0){
125                     link.put(s[0],d);
126                 }
127             }
128         }
129         double sum=0;
130         for(Map.Entry<String,Double> entry:alaph.entrySet()){
131             sum+=link.get(entry.getKey())*entry.getValue();
132         }
133         
134         System.out.println("    ");
135         System.out.println("sum is "+sum);
136         System.out.println("    ");
137         double result=factor*sum+(1-factor)/num;
138         context.write(_key,new Text(String.valueOf(result)+":"+content));
139         
140     }
141 
142 }

def f(x):
    links=x[1][0]
    rank=x[1][1]
    n=len(links.split(","))
    result=[]
    for s in links.split(","):
        result.append((s,rank*1.0/n))
    return result

file="hdfs://10.107.8.110:9000/spark_test/pagerank.txt"

data=sc.textFile(file)
link=data.map(lambda x:(x.split(":")[0],x.split(":")[1]))
n=data.count()
rank=link.mapValues(lambda x:1.0/n)

for i in range(10):
    rank=link.join(rank).flatMap(f).reduceByKey(lambda x,y:x+y).mapValues(lambda x:0.15/n+0.85*x)