我正在尝试实现不同的模式:
map(key, record):
emit record,null
reduce(key, records):
emit key
我的钥匙是复杂的,定制的 Writable
. 如果我在reduce中发出键及其哈希代码:
context.write(key, new IntWtitable(key.hashCode());
我收到以下输出:
key1 -1808937256
key2 -768063202
key3 906064410
key2 -768063202
key3 906064410
理论上,输出应该只包含 key1
, key2
,和 key3
因为我用的是 HashPartitioner
:具有相等哈希代码的键组合到同一分区中。这里显然不是这样。
如果我要转换我的情结 Writable
变成一个 Text
对象(并相应地调整Map器/还原器类),并在 Mapper
:
context.write(new Text(key.toString()), NullWritable.get());
... 输出如预期:
key1 1013632023
key2 762485389
key3 -1193948769
好的,这里有一个简单的例子来说明这个行为。
输入:
A A A A A
B B B B B
C C C C C
A A A A A
B B B B B
mapreduce作业:
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
import org.apache.hadoop.conf.Configured;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.ArrayWritable;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.Writable;
import org.apache.hadoop.io.WritableComparable;
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.output.FileOutputFormat;
import org.apache.hadoop.util.Tool;
import org.apache.hadoop.util.ToolRunner;
public class DistinctPattern extends Configured implements Tool {
public static class DistinctMapper extends Mapper<Object, Text, ComplexObject, NullWritable> {
public void map(Object key, Text value, Context context) throws IOException, InterruptedException {
ComplexObject o = new ComplexObject(value.toString());
context.write(o, NullWritable.get());
}
}
public static class DistinctReducer extends Reducer<ComplexObject, NullWritable, ComplexObject, IntWritable> {
public void reduce(ComplexObject key, Iterable<NullWritable> values, Context context)
throws IOException, InterruptedException {
context.write(key, new IntWritable(key.hashCode()));
}
}
public static class MyArrayWritable extends ArrayWritable {
public MyArrayWritable(Writable[] values) {
super(DatumObject.class, values);
}
public MyArrayWritable() {
super(DatumObject.class);
}
@Override
public String toString() {
return Arrays.toString(get());
}
}
public static class DatumObject implements Writable {
private String datum;
public DatumObject() {}
public DatumObject(String d) {
datum = d;
}
@Override
public void readFields(DataInput in) throws IOException {
datum = in.readUTF();
}
@Override
public void write(DataOutput out) throws IOException {
out.writeUTF(datum);
}
@Override
public String toString() {
return datum;
}
@Override
public int hashCode() {
return 31 * datum.hashCode();
}
}
public static class ComplexObject implements WritableComparable<ComplexObject> {
private List<DatumObject> data = new ArrayList<>();
public ComplexObject() {}
public ComplexObject(String d) {
String[] elements = d.split(" ");
for(int i = 0; i < elements.length; i++)
data.add(new DatumObject(elements[i]));
}
public int size() {
return data.size();
}
@Override
public void readFields(DataInput in) throws IOException {
data.clear();
MyArrayWritable m = new MyArrayWritable();
m.readFields(in);
Writable[] w = m.get();
for(int i = 0; i < w.length; i++)
data.add((DatumObject) w[i]);
}
@Override
public void write(DataOutput out) throws IOException {
MyArrayWritable m = new MyArrayWritable(data.toArray(new DatumObject[data.size()]));
m.write(out);
}
@Override
public int compareTo(ComplexObject o) {
if(this.equals(o))
return 0;
if(o.size() < this.size())
return -1;
return 1;
}
@Override
public boolean equals(Object obj) {
if(!(obj instanceof ComplexObject))
return false;
ComplexObject other = (ComplexObject) obj;
return other.data.equals(data);
}
@Override
public int hashCode() {
return 31 * data.hashCode();
}
@Override
public String toString() {
StringBuilder s= new StringBuilder();
data.forEach( entry -> {
s.append(entry);
s.append(" ");
});
return s.toString();
}
}
@Override
public int run(String[] args) throws Exception {
Job job = Job.getInstance();
job.setJar("distinct.jar");
job.setJarByClass(DistinctPattern.class);
job.setMapperClass(DistinctMapper.class);
job.setReducerClass(DistinctReducer.class);
job.setMapOutputKeyClass(ComplexObject.class);
job.setMapOutputValueClass(NullWritable.class);
job.setOutputKeyClass(ComplexObject.class);
job.setOutputValueClass(IntWritable.class);
FileInputFormat.addInputPath(job, new Path(args[0]));
FileOutputFormat.setOutputPath(job, new Path(args[1]));
return job.waitForCompletion(true) ? 0 : 1;
}
public static void main(String[] args) throws Exception {
int exitCode = ToolRunner.run(new DistinctPattern(), args);
System.exit(exitCode);
}
}
预期产量:
A A A A A 368623362
B B B B B 1285710467
C C C C C -2092169724
实际输出:
A A A A A 368623362
B B B B B 1285710467
C C C C C -2092169724
A A A A A 368623362
B B B B B 1285710467
我错过了什么?
ps:hadoop 2.7.3版
1条答案
按热度按时间nnt7mjpx1#
好的,在我的代码中发现了错误。首先,最小工作示例缺少
equals
类中的方法DatumObject
:第二,有一个方面我无法在最小的工作示例中重现,但它出现在我的实际代码中,那就是我的
key
类确实实现了WritableComparable
接口。因此,我怀疑shuffle阶段没有按预期对密钥进行排序。一旦compareTo
方法在构成我的key
值(请参见这里的类图),不同的模式按预期工作。