Hadoop MapReduce WordPairsCount produces inconsistent results

Question

I have a very confusing results when I run MapReduce on Hadoop. Here is the code (see below). As you can see, it is a very simple MapReduce operation. The input is 1 directory with 100 .lineperdoc files (wikipedia articles represented as 1 line). We find the pairs of words in a line of text (map) and then in the Reduce function we sum up the counts for the same words pair key and output this pair if the sum>=500. The problem arises in the Reducer: if I have condition sum>=100 I get different results from when I increase the threshold to 500, 1000, etc (please see image and description below after the code).

import java.io.IOException;
import java.util.TreeMap;

import javax.naming.Context;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.conf.Configured;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
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.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
import org.apache.hadoop.util.Tool;
import org.apache.hadoop.util.ToolRunner;

public class HadoopWordPairsb2 extends Configured implements Tool {

    public static class Map extends Mapper<LongWritable, Text, Text, IntWritable> {
        private final static IntWritable one = new IntWritable(1);
        private Text pair = new Text();
        private Text lastWord = new Text();

        @Override
        public void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

            String[] splitLine = value.toString().split(" ");

            for (String w : splitLine) {
                if (lastWord.getLength() > 0) {
                    // only consider words
                    if (w.matches("[a-zA-Z]+")) {
                        pair.set(lastWord + ":" + w);
                        context.write(pair, one);
                    }
                }
                lastWord.set(w);
            }
        }
    }

    public static class Reduce extends Reducer<Text, IntWritable, Text, IntWritable> {

        @Override
        public void reduce(Text key, Iterable<IntWritable> values, Context context)
                throws IOException, InterruptedException {
            
            Integer sum = 0;

            for (IntWritable value : values)
                sum += value.get();
            
            // output only words that occured more than 500 times
            if (sum >= 500) {
                context.write(key, new IntWritable(sum));
            }
        }
    }

    @Override
    public int run(String[] args) throws Exception {
        Job job = Job.getInstance(new Configuration(), "HadoopWordPairsb2");
        job.setJarByClass(HadoopWordPairsb2.class);

        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(IntWritable.class);
        job.setMapperClass(Map.class);
        job.setCombinerClass(Reduce.class);
        job.setReducerClass(Reduce.class);

        job.setInputFormatClass(TextInputFormat.class);
        job.setOutputFormatClass(TextOutputFormat.class);

        FileInputFormat.setInputPaths(job, args[0]);
        FileOutputFormat.setOutputPath(job, new Path(args[1]));

        job.waitForCompletion(true);
        return 0;
    }

    public static void main(String[] args) throws Exception {
        int ret = ToolRunner.run(new Configuration(), new HadoopWordPairsb2(), args);
        System.exit(ret);
    }
}

In the above image, the output on the top is for the condition sum>=100. The output right after it is for sum>=500. As you can clearly see, the results are inconsistent and I wonder why is that the case. Please let me know if I miss something very obvious, otherwise I have no idea... Thanks!

upd1: changed Map function to this:

@Override
    public void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

      String line = value.toString().toLowerCase();
      String[] tokens = line.split("[^a-z0-9_.]+");
      for (String token : tokens) {
        if (lastWord.getLength() > 0) {
          pair.set(lastWord + ":" + token);
          context.write(pair, one);
        }
        lastWord.set(token);
      }
    }

and got these results (for sum>=1000 and sum>=500, respectively). Still inconsistent...

Answer 1

SOLUTION: by removing the combiner class, the reducer works just fine. I guess there was something wrong with it, but I cannot say what exactly.