java - Hadoop Map Reduce - 将 Iterable 值写入上下文时，reduce 中的嵌套循环忽略文本结果

Question

我是 hadoop 的新手，我试图在一个简单的输入文件上运行 map reduce(参见示例)。
我尝试使用两个 for 循环从属性列表中制作某种笛卡尔积，并且由于某种原因，我得到的结果值始终为空。
我试图用它来调整它，最终它只有在我在迭代它时设置结果 Text 时才起作用(我知道，这对我来说也很奇怪)。
如果您能帮助我理解问题，我将不胜感激，可能是我做错了什么。

这是我拥有的输入文件。

A 1
B 2
C 1
D 2
C 2
E 1

我想得到以下输出:

1 A-C, A-E, C-E
2 B-C, B-D, C-D

所以我尝试实现以下map reduce类:
公共(public)类 DigitToPairOfLetters {

    public static class TokenizerMapper
            extends Mapper<Object, Text, Text, Text> {

        private Text digit = new Text();
        private Text letter = new Text();

        public void map(Object key, Text value, Context context
                ) throws IOException, InterruptedException {
            StringTokenizer itr = new StringTokenizer(value.toString());
            while (itr.hasMoreTokens()) {
                letter.set(itr.nextToken());
                digit.set(itr.nextToken());
                context.write(digit, letter);
            }
        }
    }

    public static class DigitToLetterReducer
            extends Reducer<Text, Text, Text, Text> {
        private Text result = new Text();

        public void reduce(Text key, Iterable<Text> values,
                Context context
                ) throws IOException, InterruptedException {
            List<String> valuesList = new ArrayList<>();
            for (Text value :values) {
                valuesList.add(value.toString());
            }
            StringBuilder builder = new StringBuilder();
            for (int i=0; i<valuesList.size(); i++) {
                for (int j=i+1; j<valuesList.size(); j++) {
                    builder.append(valuesList.get(i)).append(" 
").append(valuesList.get(j)).append(",");
                }
            }
            context.write(key, result);
        }
    }

    public static void main(String[] args) throws Exception {
        Configuration conf = new Configuration();
        Job job = Job.getInstance(conf, "digit to letter");
        job.setJarByClass(DigitToPairOfLetters.class);
        job.setMapperClass(TokenizerMapper.class);
        job.setCombinerClass(DigitToLetterReducer.class);
        job.setReducerClass(DigitToLetterReducer.class);
        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(Text.class);
        FileInputFormat.addInputPath(job, new Path(args[0]));
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        System.exit(job.waitForCompletion(true) ? 0 : 1);
    }
}

但是这段代码将为我提供以下空列表的输出:

1
2

当我在 for 循环中添加结果集时，它似乎可以工作:
公共(public)类 DigitToPairOfLetters {

    public static class TokenizerMapper
            extends Mapper<Object, Text, Text, Text> {

        private Text digit = new Text();
        private Text letter = new Text();

        public void map(Object key, Text value, Context context
                ) throws IOException, InterruptedException {
            StringTokenizer itr = new StringTokenizer(value.toString());
            while (itr.hasMoreTokens()) {
                letter.set(itr.nextToken());
                digit.set(itr.nextToken());
                context.write(digit, letter);
            }
        }
    }

    public static class DigitToLetterReducer
            extends Reducer<Text, Text, Text, Text> {
        private Text result = new Text();

        public void reduce(Text key, Iterable<Text> values,
                Context context
                ) throws IOException, InterruptedException {
            List<String> valuesList = new ArrayList<>();
            for (Text value :values) {
                valuesList.add(value.toString());
                // TODO: We set the valuesList in the result since otherwise the 
hadoop process will ignore the values
                // in it.
                result.set(valuesList.toString());
            }
            StringBuilder builder = new StringBuilder();
            for (int i=0; i<valuesList.size(); i++) {
                for (int j=i+1; j<valuesList.size(); j++) {
                    builder.append(valuesList.get(i)).append(" 
").append(valuesList.get(j)).append(",");
                    // TODO: We set the builder every iteration in the loop since otherwise the hadoop process will
                    // ignore the values
                    result.set(builder.toString());
                }
            }
            context.write(key, result);
        }
    }

    public static void main(String[] args) throws Exception {
        Configuration conf = new Configuration();
        Job job = Job.getInstance(conf, "digit to letter");
        job.setJarByClass(DigitToPairOfLetters.class);
        job.setMapperClass(TokenizerMapper.class);
        job.setCombinerClass(DigitToLetterReducer.class);
        job.setReducerClass(DigitToLetterReducer.class);
        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(Text.class);
        FileInputFormat.addInputPath(job, new Path(args[0]));
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        System.exit(job.waitForCompletion(true) ? 0 : 1);
    }
}

这会给我以下结果:

1   [A C,A E,C E]
2   [B C,B D,C D]

我会很感激你的帮助

Answer 1

您的第一种方法似乎很好，您只需要添加以下行:

result.set(builder.toString());

前

context.write(key, result);

就像你在第二个函数中所做的那样。

Context.write 刷新输出，因为 result 只是一个空对象，没有任何值作为值传递，只有键被传递。因此，在传递之前，您需要将值(A-E 等)设置到结果中。