1. Introduction To Hadoop
Kenneth Heafield
Google Inc
January 14, 2008
Example code from Hadoop 0.13.1 used under the Apache License Version 2.0
and modified for presentation. Except as otherwise noted, the content of this
presentation is licensed under the Creative Commons Attribution 2.5 License.
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 1 / 12

2. Outline
1 Word Count Code
Mapper
Reducer
Main
2 How it Works
Serialization
Data Flow
3 Lab
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 2 / 12

3. Word Count Code Mapper
Mapper
< “wikipedia.org”, “The Free” >→ < “The”, 1 >, < “Free”, 1 >
public void map(WritableComparable key,
Writable value, OutputCollector output,
Reporter reporter) throws IOException {
String line = ((Text)value).toString();
StringTokenizer itr = new StringTokenizer(line);
Text word = new Text();
while (itr.hasMoreTokens()) {
word.set(itr.nextToken());
output.collect(word, new IntWritable(1));
}
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 3 / 12

4. Word Count Code Mapper
Mapper
< “wikipedia.org”, “The Free” >→ < “The”, 1 >, < “Free”, 1 >
public void map(WritableComparable key,
Writable value, OutputCollector output,
Reporter reporter) throws IOException {
String line = ((Text)value).toString();
StringTokenizer itr = new StringTokenizer(line);
Text word = new Text();
while (itr.hasMoreTokens()) {
word.set(itr.nextToken());
output.collect(word, new IntWritable(1));
}
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 3 / 12

5. Word Count Code Mapper
Mapper
< “wikipedia.org”, “The Free” >→ < “The”, 1 >, < “Free”, 1 >
public void map(WritableComparable key,
Writable value, OutputCollector output,
Reporter reporter) throws IOException {
String line = ((Text)value).toString();
StringTokenizer itr = new StringTokenizer(line);
Text word = new Text();
while (itr.hasMoreTokens()) {
word.set(itr.nextToken());
output.collect(word, new IntWritable(1));
}
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 3 / 12

6. Word Count Code Mapper
Mapper
< “wikipedia.org”, “The Free” >→ < “The”, 1 >, < “Free”, 1 >
public void map(WritableComparable key,
Writable value, OutputCollector output,
Reporter reporter) throws IOException {
String line = ((Text)value).toString();
StringTokenizer itr = new StringTokenizer(line);
Text word = new Text();
while (itr.hasMoreTokens()) {
word.set(itr.nextToken());
output.collect(word, new IntWritable(1));
}
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 3 / 12

7. Word Count Code Mapper
Mapper
< “wikipedia.org”, “The Free” >→ < “The”, 1 >, < “Free”, 1 >
public void map(WritableComparable key,
Writable value, OutputCollector output,
Reporter reporter) throws IOException {
String line = ((Text)value).toString();
StringTokenizer itr = new StringTokenizer(line);
Text word = new Text();
while (itr.hasMoreTokens()) {
word.set(itr.nextToken());
output.collect(word, new IntWritable(1));
}
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 3 / 12

8. Word Count Code Reducer
Reducer
< “The”, 1 >, < “The”, 1 >→ < “The”, 2 >
public void reduce(WritableComparable key,
Iterator values,
OutputCollector output,
Reporter reporter)
throws IOException {
int sum = 0;
while (values.hasNext()) {
sum += ((IntWritable) values.next()).get();
}
output.collect(key, new IntWritable(sum));
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 4 / 12

9. Word Count Code Reducer
Reducer
< “The”, 1 >, < “The”, 1 >→ < “The”, 2 >
public void reduce(WritableComparable key,
Iterator values,
OutputCollector output,
Reporter reporter)
throws IOException {
int sum = 0;
while (values.hasNext()) {
sum += ((IntWritable) values.next()).get();
}
output.collect(key, new IntWritable(sum));
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 4 / 12

10. Word Count Code Reducer
Reducer
< “The”, 1 >, < “The”, 1 >→ < “The”, 2 >
public void reduce(WritableComparable key,
Iterator values,
OutputCollector output,
Reporter reporter)
throws IOException {
int sum = 0;
while (values.hasNext()) {
sum += ((IntWritable) values.next()).get();
}
output.collect(key, new IntWritable(sum));
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 4 / 12

11. Word Count Code Reducer
Reducer
< “The”, 1 >, < “The”, 1 >→ < “The”, 2 >
public void reduce(WritableComparable key,
Iterator values,
OutputCollector output,
Reporter reporter)
throws IOException {
int sum = 0;
while (values.hasNext()) {
sum += ((IntWritable) values.next()).get();
}
output.collect(key, new IntWritable(sum));
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 4 / 12

12. Word Count Code Main
Main
public static void main(String[] args)
throws IOException {
JobConf conf = new JobConf(WordCount.class);
conf.setJobName("wordcount");
conf.setMapperClass(MapClass.class);
conf.setCombinerClass(ReduceClass.class);
conf.setReducerClass(ReduceClass.class);
conf.setNumMapTasks(new Integer(40));
conf.setNumReduceTasks(new Integer(30));
conf.setInputPath(new Path("/shared/wikipedia_small"));
conf.setOutputPath(new Path("/user/kheafield/word_count"));
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);
JobClient.runJob(conf);
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 5 / 12

13. Word Count Code Main
Main
public static void main(String[] args)
throws IOException {
JobConf conf = new JobConf(WordCount.class);
conf.setJobName("wordcount");
conf.setMapperClass(MapClass.class);
conf.setCombinerClass(ReduceClass.class);
conf.setReducerClass(ReduceClass.class);
conf.setNumMapTasks(new Integer(40));
conf.setNumReduceTasks(new Integer(30));
conf.setInputPath(new Path("/shared/wikipedia_small"));
conf.setOutputPath(new Path("/user/kheafield/word_count"));
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);
JobClient.runJob(conf);
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 5 / 12

14. Word Count Code Main
Main
public static void main(String[] args)
throws IOException {
JobConf conf = new JobConf(WordCount.class);
conf.setJobName("wordcount");
conf.setMapperClass(MapClass.class);
conf.setCombinerClass(ReduceClass.class);
conf.setReducerClass(ReduceClass.class);
conf.setNumMapTasks(new Integer(40));
conf.setNumReduceTasks(new Integer(30));
conf.setInputPath(new Path("/shared/wikipedia_small"));
conf.setOutputPath(new Path("/user/kheafield/word_count"));
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);
JobClient.runJob(conf);
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 5 / 12

15. Word Count Code Main
Main
public static void main(String[] args)
throws IOException {
JobConf conf = new JobConf(WordCount.class);
conf.setJobName("wordcount");
conf.setMapperClass(MapClass.class);
conf.setCombinerClass(ReduceClass.class);
conf.setReducerClass(ReduceClass.class);
conf.setNumMapTasks(new Integer(40));
conf.setNumReduceTasks(new Integer(30));
conf.setInputPath(new Path("/shared/wikipedia_small"));
conf.setOutputPath(new Path("/user/kheafield/word_count"));
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);
JobClient.runJob(conf);
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 5 / 12

16. How it Works Serialization
Types
Purpose
Simple serialization for keys, values, and other data
Interface Writable
Read and write binary format
Convert to String for text formats
WritableComparable adds sorting order for keys
Example Implementations
ArrayWritable is only Writable
BooleanWritable
IntWritable sorts in increasing order
Text holds a String
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 6 / 12

17. How it Works Serialization
A Writable
public class IntPairWritable implements Writable {
public int first;
public int second;
public void write(DataOutput out) throws IOException {
out.writeInt(first);
out.writeInt(second);
}
public void readFields(DataInput in) throws IOException {
first = in.readInt();
second = in.readInt();
}
public int hashCode() { return first + second; }
public String toString() {
return Integer.toString(first) + "," +
Integer.toString(second);
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 7 / 12

18. How it Works Serialization
WritableComparable Method
public int compareTo(Object other) {
IntPairWritable o = (IntPairWritable)other;
if (first < o.first) return -1;
if (first > o.first) return 1;
if (second < o.second) return -1;
if (second > o.second) return 1;
return 0;
}
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 8 / 12

19. How it Works Data Flow
Data Flow
Default Flow
1 Mappers read from HDFS
2 Map output is partitioned by key and sent to Reducers
3 Reducers sort input by key
4 Reduce output is written to HDFS
1 HDFS 2 Mapper 3 Reducer 4 HDFS
Input Map Sort Reduce Output
Input Map Sort Reduce Output
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 9 / 12

20. How it Works Data Flow
Combiners
Concept
Add counts at Mapper before sending to Reducer.
Word count is 6 minutes with combiners and 14 without.
Implementation
Mapper caches output and periodically calls Combiner
Input to Combine may be from Map or Combine
Combiner uses interface as Reducer
Mapper
Combine Sort Reduce Output
Input Map
Cache Sort Reduce Output
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 10 / 12

21. Lab
Exercises
Recommended: Word Count
Get word count running.
Bigrams
Count bigrams and unigrams efficiently.
Capitalization
With what probability is a word capitalized?
Indexer
In what documents does each word appear? Where in the documents?
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 11 / 12

22. Lab
Instructions
1 Login to the cluster successfully (and set your password).
2 Get Eclipse installed, so you can build Java code.
3 Install the Hadoop plugin for Eclipse so you can deploy jobs to the
cluster.
4 Set up your Eclipse workspace from a template that we provide.
5 Run the word counter example over the Wikipedia data set.
Kenneth Heafield (Google Inc) Introduction To Hadoop January 14, 2008 12 / 12