hadoop处理前N个最值问题

例子为100W 条数据 取出前十个最值（纯本人看完课程后的手写，没有参考网上，结果应该没问题的，也没找到标准答案写法。。）

首先，由于值都是double，默认的排序方式是升序，这里面我们取得是降序，所以自定义hadoop对象，并实现WritableComparable接口，然后覆盖compareTo方法。

class MySuperKey implements WritableComparable<MySuperKey>{
	Long mykey;
	public MySuperKey(){

	}

	public MySuperKey(long mykey){
		this.mykey=mykey;
	}
	@Override
	public void readFields(DataInput in) throws IOException {
		this.mykey=in.readLong();
	}

	@Override
	public void write(DataOutput out) throws IOException {
		out.writeLong(this.mykey);
	}



	@Override
	public int hashCode() {
		// TODO Auto-generated method stub
		return this.mykey.hashCode();
	}

	@Override
	public boolean equals(Object obj) {
		if(! (obj instanceof LongWritable)){
			return false;
		}
		LongWritable lvv=(LongWritable)obj;
		return (this.mykey==lvv.get());
	}

	@Override
	public int compareTo(MySuperKey o) {	
		return ((int)(o.mykey-this.mykey));
	}
}

由于在map函数中执行后 需要对相同的key值进行分组，但对于自己创建的对象，无法判断是否是相同的，hadoop基础类型是可以的，此时，需要实现RawComparator接口，并覆盖compare方法，并在job执行的时候，加上

job.setGroupingComparatorClass(MyGroupingComparator.class);

下面是自定义的分组对象

class MyGroupingComparator implements RawComparator<MySuperKey>{

	@Override
	public int compare(MySuperKey o1, MySuperKey o2) {
		return (int)(o1.mykey-o2.mykey);
	}

	@Override
	public int compare(byte[] arg0, int arg1, int arg2, byte[] arg3, int arg4,
			int arg5) {
		return WritableComparator.compareBytes(arg0, arg1, 8, arg3, arg4, 8);
	}	
}

下面覆盖map和reduce方法

class mysuperMap extends Mapper<LongWritable,Text,MySuperKey,NullWritable>{
	protected void map(LongWritable key, Text value, org.apache.hadoop.mapreduce.Mapper<LongWritable,Text,MySuperKey,NullWritable>.Context context) throws IOException ,InterruptedException {
		long sdsd=Long.parseLong(value.toString());
		MySuperKey my=new MySuperKey(sdsd);
		context.write(my,NullWritable.get());
	};
}

class mysupderreduace extends Reducer<MySuperKey, NullWritable, LongWritable, NullWritable>{
	int i=0;
	protected void reduce(MySuperKey key, java.lang.Iterable<NullWritable> value, org.apache.hadoop.mapreduce.Reducer<MySuperKey,NullWritable,LongWritable,NullWritable>.Context arg2) throws IOException ,InterruptedException {
		i=i+1;
		if(i<11){
			arg2.write(new LongWritable(key.mykey), NullWritable.get());
		}
	};
}

下面写main函数 ，执行job

public static void main(String[] args) throws Exception {
		final String INPUT_PATHs = "hdfs://chaoren:9000/seq100w.txt";
		final String OUT_PATHs = "hdfs://chaoren:9000/out";
		Job job=new Job(new Configuration(),MySuper.class.getSimpleName());
		FileInputFormat.setInputPaths(job, INPUT_PATHs);
		job.setInputFormatClass(TextInputFormat.class);

		job.setMapperClass(mysuperMap.class);
		job.setMapOutputKeyClass(MySuperKey.class);
		job.setMapOutputValueClass(NullWritable.class);

		//1.3 指定分区类
		job.setPartitionerClass(HashPartitioner.class);
		job.setNumReduceTasks(1);


		//指定分组
		job.setGroupingComparatorClass(MyGroupingComparator.class);

		job.setReducerClass(mysupderreduace.class);
		job.setOutputKeyClass(LongWritable.class);
		job.setOutputValueClass(NullWritable.class);

		FileOutputFormat.setOutputPath(job,new Path(OUT_PATHs));
		job.setOutputFormatClass(TextOutputFormat.class);

		job.waitForCompletion(true);
	}