我决定创建自己的WritableComparable类来学习Hadoop是如何使用它的。因此,我创建了一个带有两个实例变量的Order类(orderNumber客户机),并实现了所需的方法。我还使用Eclipse生成器生成getter/setter/hashCode/equals/toString。
在compareTo中,我决定只使用orderNumber变量。
我创建了一个简单的MapReduce作业,仅用于统计数据集中某个顺序的出现次数。由于错误,我的一个测试记录是Ita而不是it
123 Ita
123 Itá
123 Itá
345 Carol
345 Carol
345 Carol
345 Carol
456 Iza Smith
据我所知,第一条记录应被视为不同的顺序,因为记录1 hashCode与记录2和3 hashCode不同。
但在reduce阶段,3条记录被分组在一起。如你所见:
Order [cliente=Ita, orderNumber=123] 3
Order [cliente=Carol, orderNumber=345] 4
Order [cliente=Iza Smith, orderNumber=456] 1
我认为它应该有一行用来存放计数为2的记录,而它应该有计数为1的记录。
因为我在compareTo中只使用了orderNumber,所以我尝试在此方法中使用String客户端(在下面的代码中注释)。然后,它就像我期望的那样工作了。
那么,这是预期的结果吗?hadoop不应该只使用hashCode来分组键及其值吗?
下面是Order类(我提交了getter/setter):public class Order implements WritableComparable<Order>
{
private String cliente;
private long orderNumber;
@Override
public void readFields(DataInput in) throws IOException
{
cliente = in.readUTF();
orderNumber = in.readLong();
}
@Override
public void write(DataOutput out) throws IOException
{
out.writeUTF(cliente);
out.writeLong(orderNumber);
}
@Override
public int compareTo(Order o) {
long thisValue = this.orderNumber;
long thatValue = o.orderNumber;
return (thisValue < thatValue ? -1 :(thisValue == thatValue ? 0 :1));
//return this.cliente.compareTo(o.cliente);
}
@Override
public int hashCode() {
final int prime = 31;
int result = 1;
result = prime * result + ((cliente == null) ? 0 : cliente.hashCode());
result = prime * result + (int) (orderNumber ^ (orderNumber >>> 32));
return result;
}
@Override
public boolean equals(Object obj) {
if (this == obj)
return true;
if (obj == null)
return false;
if (getClass() != obj.getClass())
return false;
Order other = (Order) obj;
if (cliente == null) {
if (other.cliente != null)
return false;
} else if (!cliente.equals(other.cliente))
return false;
if (orderNumber != other.orderNumber)
return false;
return true;
}
@Override
public String toString() {
return "Order [cliente=" + cliente + ", orderNumber=" + orderNumber + "]";
}
下面是MapReduce的代码:
public class TesteCustomClass extends Configured implements Tool
{
public static class Map extends MapReduceBase implements Mapper<LongWritable, Text, Order, LongWritable>
{
LongWritable outputValue = new LongWritable();
String[] campos;
Order order = new Order();
@Override
public void configure(JobConf job)
{
}
@Override
public void map(LongWritable key, Text value, OutputCollector<Order, LongWritable> output, Reporter reporter) throws IOException
{
campos = value.toString().split("t");
order.setOrderNumber(Long.parseLong(campos[0]));
order.setCliente(campos[1]);
outputValue.set(1L);
output.collect(order, outputValue);
}
}
public static class Reduce extends MapReduceBase implements Reducer<Order, LongWritable, Order,LongWritable>
{
@Override
public void reduce(Order key, Iterator<LongWritable> values,OutputCollector<Order,LongWritable> output, Reporter reporter) throws IOException
{
LongWritable value = new LongWritable(0);
while (values.hasNext())
{
value.set(value.get() + values.next().get());
}
output.collect(key, value);
}
}
@Override
public int run(String[] args) throws Exception {
JobConf conf = new JobConf(getConf(),TesteCustomClass.class);
conf.setMapperClass(Map.class);
// conf.setCombinerClass(Reduce.class);
conf.setReducerClass(Reduce.class);
conf.setJobName("Teste - Custom Classes");
conf.setOutputKeyClass(Order.class);
conf.setOutputValueClass(LongWritable.class);
conf.setInputFormat(TextInputFormat.class);
conf.setOutputFormat(TextOutputFormat.class);
FileInputFormat.setInputPaths(conf, new Path(args[0]));
FileOutputFormat.setOutputPath(conf, new Path(args[1]));
JobClient.runJob(conf);
return 0;
}
public static void main(String[] args) throws Exception {
int res = ToolRunner.run(new Configuration(),new TesteCustomClass(),args);
System.exit(res);
}
}
默认分区器是HashPartitioner
,它使用hashCode
方法来确定将K,V对发送到哪个reducer。
一旦在reducer中(或者如果你使用的是在map端运行的Combiner), compareTo
方法用于对键进行排序,然后也使用(默认情况下)来比较顺序键是否应该分组在一起,以及它们的关联值是否应该在同一迭代中减少。
如果你不使用cliente
键变量,只有你的orderNumber
变量在你的compareTo
方法,那么任何键与相同的orderNumber
将有其值减少在一起-不管cliente
值(这是你目前正在观察的)