hadoop的分区学习

自定义分区实现：

抽象类

要自定义分区规则，就必须继承并且重写。

设置分区数量

driver类：

        job.setNumReduceTasks(3);
        job.setPartitionerClass(FlowPartitioner.class);

分区方法类，直接用编号代替就好了

public class FlowPartitioner extends Partitioner<Text, FlowBean> {
    @Override
    public int getPartition(Text text, FlowBean flowBean, int numPartitions) {
        int partitions;
        String phoneNum = text.toString();
        if(phoneNum.startsWith("136")){
            partitions = 0;
        } else if (phoneNum.startsWith("137")) {
            partitions = 1;
        }else if (phoneNum.startsWith("138")) {
            partitions = 2;
        }else if (phoneNum.startsWith("139")) {
            partitions = 3;
        }else {
            partitions = 4;
        }
        return partitions;
    }
}