hadoop mapreducez自定義分割槽

hgs19921112發表於2018-09-03
package hello_hadoop;
import java.io.IOException;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.DoubleWritable;
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.Partitioner;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
public class AutoParitionner {
	public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {
		if(args.length!=2) {
			System.err.println("Usage: hadoop jar xxx.jar <input path> <output path>");
			System.exit(1);
		}
		Configuration conf = new Configuration();
		Job job = Job.getInstance(conf, "avg of grades");
		job.setJarByClass(AutoParitionner.class);
		job.setMapperClass(PartitionInputClass.class);
		job.setReducerClass(PartitionOutputClass.class);
		job.setMapOutputKeyClass(Text.class);
		job.setMapOutputValueClass(DoubleWritable.class);
		job.setOutputKeyClass(Text.class);
		job.setOutputValueClass(DoubleWritable.class);
		//宣告自定義分割槽的類,下面有類的宣告
		job.setPartitionerClass(MyPartitioner.class);
		job.setNumReduceTasks(2);
		FileInputFormat.addInputPath(job, new Path(args[0]));
		FileOutputFormat.setOutputPath(job, new Path(args[1]));
		
		System.exit(job.waitForCompletion(true)?0:1);
		
	}
	}
class PartitionInputClass extends Mapper<LongWritable, Text, Text, DoubleWritable>{
	@Override
	protected void map(LongWritable key, Text value, Mapper<LongWritable, Text, Text, DoubleWritable>.Context context)
			throws IOException, InterruptedException {
		String line = value.toString();
		if(line.length()>0){
			String[] array = line.split("\t");
			if(array.length==2){
				String name=array[0];
				int grade = Integer.parseInt(array[1]);
				context.write(new Text(name), new DoubleWritable(grade));
			}
		}
		
		
	}
	
}
class PartitionOutputClass extends Reducer<Text, DoubleWritable, Text, DoubleWritable>{
	@Override
	protected void reduce(Text text, Iterable<DoubleWritable> iterable,
			Reducer<Text, DoubleWritable, Text, DoubleWritable>.Context context) throws IOException, InterruptedException {
		int sum = 0;
		int cnt= 0 ;
		for(DoubleWritable iw : iterable) {
			sum+=iw.get();
			cnt++;
		}
		context.write(text, new DoubleWritable(sum/cnt));
	}
	
}
//自定義分割槽的類
//Partitioner<Text  , DoubleWritable > Text,DoubleWirtable分別為map結果的key,value
class MyPartitioner extends Partitioner<Text  , DoubleWritable >{
	@Override
	public int getPartition(Text text, DoubleWritable value, int numofreuceTask) {
		String name = text.toString();
		if(name.equals("wd")||name.equals("wzf")||name.equals("xzh")||name.equals("zz")) {
			return 0;		
		}else
			return 1;
	}
	}


來自 “ ITPUB部落格 ” ,連結:http://blog.itpub.net/31506529/viewspace-2213415/,如需轉載,請註明出處,否則將追究法律責任。

相關文章