java - cleanup(context) 方法有什么作用？

Question

我不明白 Hadoop 中的清理方法到底是做什么的，它是如何工作的？我有以下 Map-Reduce 代码来计算一堆数字的最大值、最小值和平均值。

public class Statistics 
{
    public static class Map extends Mapper<LongWritable, Text, Text, Text>
    {
        public void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException 
        {
            /* code to calculate min, max, and mean from among a bunch of numbers */
        }
        public void cleanup(Context context) throws IOException, InterruptedException
        {
            Text key_min = new Text();
            key_min.set("min");
            Text value_min = new Text();
            value_min.set(String.valueOf(min));
            context.write(key_min,value_min);

            Text key_max = new Text();
            key_max.set("max");
            Text value_max = new Text();
            value_max.set(String.valueOf(max));
            context.write(key_max,value_max);

            Text key_avg = new Text();
            key_avg.set("avg");
            Text value_avg = new Text();
            value_avg.set(String.valueOf(linear_sum)+","+count);
            context.write(key_avg,value_avg);

            Text key_stddev = new Text();
            key_stddev.set("stddev");
            Text value_stddev = new Text();
            value_stddev.set(String.valueOf(linear_sum)+","+count+","+String.valueOf(quadratic_sum));
            context.write(key_stddev,value_stddev);
        }
    }
    public static class Reduce extends Reducer<Text,Text,Text,Text>
    {
        public void reduce(Text key, Iterable<Text> values,Context context) throws IOException, InterruptedException 
        {
            /* code to further find min, max and mean from among the outputs of different mappers */
        }
    }
    public static void main(String[] args) throws Exception 
    {
        /* driver program */
    }
}

那么 cleanup(Context context) 方法到底在做什么呢？我假设它从一堆映射器收集输出(键，值)对并将其传递给缩减器。在其他网站上，我读到在 MapReduce 中运行的顺序是:setup -> map -> cleanup 然后是 setup -> reduce -> cleanup。为什么这个程序没有使用设置方法？

Answer 1

这些值必须不是在 Mapper 中计算的，它必须在 Reduce 步骤中计算。 https://hadoop.apache.org/docs/current/hadoop-mapreduce-client/hadoop-mapreduce-client-core/MapReduceTutorial.html#Reducer