0% found this document useful (0 votes)

529 views550 pages

Cloudera Developer Training

Uploaded by

Naga

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

529 views550 pages

Cloudera Developer Training

Uploaded by

Naga

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 550

Cloudera Developer Training

for Apache Hadoop

201202$

01-1

Chapter 1
Introduction

01-2

Introduction
About This Course
About Cloudera
Course Logistics

01-3

Course Objectives
During this course, you will learn:
The core technologies of Hadoop
How HDFS and MapReduce work
What other projects exist in the Hadoop ecosystem
How to develop MapReduce jobs
How Hadoop integrates into the datacenter
Algorithms for common MapReduce tasks
How to create large workflows using Oozie

01-4

Course Objectives (contd)

How Hive and Pig can be used for rapid application development
Best practices for developing and debugging MapReduce jobs
Advanced features of the Hadoop API
How to handle graph manipulation problems with MapReduce

01-5

Course Chapters
Introduction
The Motivation For Hadoop
Hadoop: Basic Contents
Writing a MapReduce Program
Integrating Hadoop Into The Workflow
Delving Deeper Into The Hadoop API
Common MapReduce Algorithms
An Introduction to Hive and Pig
Practical Development Tips and Techniques
More Advanced MapReduce Programming
Joining Data Sets in MapReduce Jobs
Graph Manipulation In MapReduce
An Introduction to Oozie
Conclusion
Appendix: Cloudera Enterprise
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

01-6

Introduction
About This Course
About Cloudera
Course Logistics

01-7

About Cloudera
Cloudera is The commercial Hadoop company
Founded by leading experts on Hadoop from Facebook, Google,
Oracle and Yahoo
Provides consulting and training services for Hadoop users
Staff includes committers to virtually all Hadoop projects

01-8

Cloudera Software
Clouderas Distribution including Apache Hadoop (CDH)
A single, easy-to-install package from the Apache Hadoop core
repository
Includes a stable version of Hadoop, plus critical bug fixes and
solid new features from the development version
100% open source
Cloudera Manager, Free Edition
The easiest way to deploy a Hadoop cluster
Automates installation of Hadoop software
Installation, monitoring and configuration is performed from a
central machine
Manages up to 50 nodes
Completely free
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

01-9

Cloudera Enterprise
Cloudera Enterprise
Complete package of software and support
Built on top of CDH
Includes full version of Cloudera Manager
Install, manage, and maintain a cluster of any size
LDAP integration
Includes powerful cluster monitoring and auditing tools
Resource consumption tracking
Proactive health checks
Alerting
Configuration change audit trails
And more
24 x 7 support
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

01-10

Cloudera Services
Provides consultancy services to many key users of Hadoop
Including AOL Advertising, Samsung, Groupon, NAVTEQ, Trulia,
Tynt, RapLeaf, Explorys Medical
Solutions Architects are experts in Hadoop and related
technologies
Several are committers to the Apache Hadoop project
Provides training in key areas of Hadoop administration and
development
Courses include System Administrator training, Developer
training, Hive and Pig training, HBase Training, Essentials for
Managers
Custom course development available
Both public and on-site training available
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

01-11

Introduction
About This Course
About Cloudera
Course Logistics

01-12

Logistics
Course start and end times
Lunch
Breaks
Restrooms
Can I come in early/stay late?
Certification

01-13

Introductions
About your instructor
About you
Experience with Hadoop?
Experience as a developer?
What programming languages do you use?
Expectations from the course?

01-14

Chapter 2
The Motivation For Hadoop

02-1

02-2

The Motivation For Hadoop

In this chapter you will learn
What problems exist with traditional large-scale computing
systems
What requirements an alternative approach should have
How Hadoop addresses those requirements

02-3

The Motivation For Hadoop

Problems with Traditional Large-Scale Systems
Requirements for a New Approach
Hadoop!
Conclusion

02-4

Traditional Large-Scale Computation

Traditionally, computation has been processor-bound
Relatively small amounts of data
Significant amount of complex processing performed on that data
For decades, the primary push was to increase the computing
power of a single machine
Faster processor, more RAM
Distributed systems evolved to allow
developers to use multiple machines
for a single job
MPI
PVM
Condor
MPI: Message Passing Interface
PVM: Parallel Virtual Machine
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

02-5

Distributed Systems: Problems

Programming for traditional distributed systems is complex
Data exchange requires synchronization
Finite bandwidth is available
Temporal dependencies are complicated
It is difficult to deal with partial failures of the system
Ken Arnold, CORBA designer:
Failure is the defining difference between distributed and local
programming, so you have to design distributed systems with the
expectation of failure
Developers spend more time designing for failure than they
do actually working on the problem itself

CORBA: Common Object Request Broker Architecture

02-6

Distributed Systems: Data Storage

Typically, data for a distributed system is stored on a SAN
At compute time, data is copied to the compute nodes
Fine for relatively limited amounts of data

SAN: Storage Area Network

02-7

The Data-Driven World

Modern systems have to deal with far more data than was the
case in the past
Organizations are generating huge amounts of data
That data has inherent value, and cannot be discarded
Examples:
Facebook over 70PB of data
eBay over 5PB of data
Many organizations are generating data at a rate of terabytes per
day

02-8

Data Becomes the Bottleneck

Moores Law has held firm for over 40 years
Processing power doubles every two years
Processing speed is no longer the problem
Getting the data to the processors becomes the bottleneck
Quick calculation
Typical disk data transfer rate: 75MB/sec
Time taken to transfer 100GB of data to the processor: approx 22
minutes!
Assuming sustained reads
Actual time will be worse, since most servers have less than
100GB of RAM available
A new approach is needed
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

02-9

The Motivation For Hadoop

Problems with Traditional Large-Scale Systems
Requirements for a New Approach
Hadoop!
Conclusion

02-10

Partial Failure Support

The system must support partial failure
Failure of a component should result in a graceful degradation of
application performance
Not complete failure of the entire system

02-11

Data Recoverability
If a component of the system fails, its workload should be
assumed by still-functioning units in the system
Failure should not result in the loss of any data

02-12

Component Recovery
If a component of the system fails and then recovers, it should
be able to rejoin the system
Without requiring a full restart of the entire system

02-13

Consistency
Component failures during execution of a job should not affect
the outcome of the job

02-14

Scalability
Adding load to the system should result in a graceful decline in
performance of individual jobs
Not failure of the system
Increasing resources should support a proportional increase in
load capacity

02-15

The Motivation For Hadoop

Problems with Traditional Large-Scale Systems
Requirements for a New Approach
Hadoop!
Conclusion

02-16

Hadoops History
Hadoop is based on work done by Google in the late 1990s/early
2000s
Specifically, on papers describing the Google File System (GFS)
published in 2003, and MapReduce published in 2004
This work takes a radical new approach to the problem of
distributed computing
Meets all the requirements we have for reliability and scalability
Core concept: distribute the data as it is initially stored in the
system
Individual nodes can work on data local to those nodes
No data transfer over the network is required for initial
processing

02-17

Core Hadoop Concepts

Applications are written in high-level code
Developers need not worry about network programming, temporal
dependencies or low-level infrastructure
Nodes talk to each other as little as possible
Developers should not write code which communicates between
nodes
Shared nothing architecture
Data is spread among machines in advance
Computation happens where the data is stored, wherever
possible
Data is replicated multiple times on the system for increased
availability and reliability

02-18

Hadoop: Very High-Level Overview

When data is loaded into the system, it is split into blocks
Typically 64MB or 128MB
Map tasks (the first part of the MapReduce system) work on
relatively small portions of data
Typically a single block
A master program allocates work to nodes such that a Map task
will work on a block of data stored locally on that node whenever
possible
Many nodes work in parallel, each on their own part of the overall
dataset

02-19

Fault Tolerance
If a node fails, the master will detect that failure and re-assign the
work to a different node on the system
Restarting a task does not require communication with nodes
working on other portions of the data
If a failed node restarts, it is automatically added back to the
system and assigned new tasks
If a node appears to be running slowly, the master can
redundantly execute another instance of the same task
Results from the first to finish will be used
Known as speculative execution

02-20

The Motivation For Hadoop

Problems with Traditional Large-Scale Systems
Requirements for a New Approach
Hadoop!
Conclusion

02-21

The Motivation For Hadoop

In this chapter you have learned
What problems exist with traditional large-scale computing
systems
What requirements an alternative approach should have
How Hadoop addresses those requirements

02-22

Chapter 3
Hadoop: Basic Concepts

03-1

03-2

Hadoop: Basic Concepts

In this chapter you will learn
What Hadoop is
What features the Hadoop Distributed File System (HDFS)
provides
The concepts behind MapReduce
How a Hadoop cluster operates
What other Hadoop Ecosystem projects exist

03-3

Hadoop: Basic Concepts

What Is Hadoop?
The Hadoop Distributed File System (HDFS)
Hands-On Exercise: Using HDFS
How MapReduce works
Hands-On Exercise: Running a MapReduce job
Anatomy of a Hadoop Cluster
Other Ecosystem Projects
Conclusion
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-4

The Hadoop Project

Hadoop is an open-source project overseen by the Apache
Software Foundation
Originally based on papers published by Google in 2003 and
2004
Hadoop committers work at several different organizations
Including Cloudera, Yahoo!, Facebook

03-5

Hadoop Components
Hadoop consists of two core components
The Hadoop Distributed File System (HDFS)
MapReduce
There are many other projects based around core Hadoop
Often referred to as the Hadoop Ecosystem
Pig, Hive, HBase, Flume, Oozie, Sqoop, etc
Many are discussed later in the course
A set of machines running HDFS and MapReduce is known as a
Hadoop Cluster
Individual machines are known as nodes
A cluster can have as few as one node, as many as several
thousands
More nodes = better performance!
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-6

Hadoop Components: HDFS

HDFS, the Hadoop Distributed File System, is responsible for
storing data on the cluster
Data is split into blocks and distributed across multiple nodes in
the cluster
Each block is typically 64MB or 128MB in size
Each block is replicated multiple times
Default is to replicate each block three times
Replicas are stored on different nodes
This ensures both reliability and availability

03-7

Hadoop Components: MapReduce

MapReduce is the system used to process data in the Hadoop
cluster
Consists of two phases: Map, and then Reduce
Between the two is a stage known as the shuffle and sort
Each Map task operates on a discrete portion of the overall
dataset
Typically one HDFS block of data
After all Maps are complete, the MapReduce system distributes
the intermediate data to nodes which perform the Reduce phase
Much more on this later!

03-8

Hadoop: Basic Concepts

03-9

HDFS Basic Concepts

HDFS is a filesystem written in Java
Based on Googles GFS
Sits on top of a native filesystem
Such as ext3, ext4 or xfs
Provides redundant storage for massive amounts of data
Using cheap, unreliable computers

03-10

HDFS Basic Concepts (contd)

HDFS performs best with a modest number of large files
Millions, rather than billions, of files
Each file typically 100MB or more
Files in HDFS are write once
No random writes to files are allowed
Append support is included in Clouderas Distribution including
Apache Hadoop (CDH) for HBase reliability
Not recommended for general use
HDFS is optimized for large, streaming reads of files
Rather than random reads

03-11

How Files Are Stored

Files are split into blocks
Each block is usually 64MB or 128MB
Data is distributed across many machines at load time
Different blocks from the same file will be stored on different
machines
This provides for efficient MapReduce processing (see later)
Blocks are replicated across multiple machines, known as
DataNodes
Default replication is three-fold
Meaning that each block exists on three different machines
A master node called the NameNode keeps track of which blocks
make up a file, and where those blocks are located
Known as the metadata
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-12

How Files Are Stored: Example

NameNode holds metadata for
the two files (Foo.txt and
Bar.txt)
DataNodes hold the actual
blocks
Each block will be 64MB or
128MB in size
Each block is replicated
three times on the cluster

03-13

More On The HDFS NameNode

The NameNode daemon must be running at all times
If the NameNode stops, the cluster becomes inaccessible
Your system administrator will take care to ensure that the
NameNode hardware is reliable!
The NameNode holds all of its metadata in RAM for fast access
It keeps a record of changes on disk for crash recovery
A separate daemon known as the Secondary NameNode takes
care of some housekeeping tasks for the NameNode
Be careful: The Secondary NameNode is not a backup
NameNode!

03-14

HDFS: Points To Note

Although files are split into 64MB or 128MB blocks, if a file is
smaller than this the full 64MB/128MB will not be used
Blocks are stored as standard files on the DataNodes, in a set of
directories specified in Hadoops configuration files
This will be set by the system administrator
Without the metadata on the NameNode, there is no way to
access the files in the HDFS cluster
When a client application wants to read a file:
It communicates with the NameNode to determine which blocks
make up the file, and which DataNodes those blocks reside on
It then communicates directly with the DataNodes to read the
data
The NameNode will not be a bottleneck
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-15

Accessing HDFS
Applications can read and write HDFS files directly via the Java
API
Covered later in the course
Typically, files are created on a local filesystem and must be
moved into HDFS
Likewise, files stored in HDFS may need to be moved to a
machines local filesystem
Access to HDFS from the command line is achieved with the
hadoop fs command

03-16

hadoop fs Examples
Copy file foo.txt from local disk to the users directory in HDFS
hadoop fs -copyFromLocal foo.txt foo.txt

This will copy the file to /user/username/foo.txt

Get a directory listing of the users home directory in HDFS
hadoop fs -ls

Get a directory listing of the HDFS root directory

hadoop fs ls /

03-17

hadoop fs Examples (contd)

Display the contents of the HDFS file /user/fred/bar.txt
hadoop fs cat /user/fred/bar.txt

Move that file to the local disk, named as baz.txt

hadoop fs copyToLocal /user/fred/bar.txt baz.txt

Create a directory called input under the users home directory

hadoop fs mkdir input

03-18

hadoop fs Examples (contd)

Delete the directory input_old and all its contents
hadoop fs rmr input_old

03-19

Hadoop: Basic Concepts

03-20

Aside: The Training Virtual Machine

During this course, you will perform numerous Hands-On
Exercises using the Cloudera Training Virtual Machine (VM)
The VM has Hadoop installed in pseudo-distributed mode
This essentially means that it is a cluster comprised of a single
node
Using a pseudo-distributed cluster is the typical way to test your
code before you run it on your full cluster
It operates almost exactly like a real cluster
A key difference is that the data replication factor is set to 1,
not 3

03-21

Hands-On Exercise: Using HDFS

In this Hands-On Exercise you will gain familiarity with
manipulating files in HDFS
Please refer to the Hands-On Exercise Manual

03-22

Hadoop: Basic Concepts

03-23

What Is MapReduce?
MapReduce is a method for distributing a task across multiple
nodes
Each node processes data stored on that node
Where possible
Consists of two phases:
Map
Reduce

03-24

Features of MapReduce
Automatic parallelization and distribution
Fault-tolerance
Status and monitoring tools
A clean abstraction for programmers
MapReduce programs are usually written in Java
Can be written in any scripting language using Hadoop
Streaming (see later)
All of Hadoop is written in Java
MapReduce abstracts all the housekeeping away from the
developer
Developer can concentrate simply on writing the Map and
Reduce functions
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-25

MapReduce: The Big Picture

03-26

MapReduce: The JobTracker

MapReduce jobs are controlled by a software daemon known as
the JobTracker
The JobTracker resides on a master node
Clients submit MapReduce jobs to the JobTracker
The JobTracker assigns Map and Reduce tasks to other nodes
on the cluster
These nodes each run a software daemon known as the
TaskTracker
The TaskTracker is responsible for actually instantiating the Map
or Reduce task, and reporting progress back to the JobTracker

03-27

MapReduce: Terminology
A job is a full program
A complete execution of Mappers and Reducers over a dataset
A task is the execution of a single Mapper or Reducer over a slice
of data
A task attempt is a particular instance of an attempt to execute a
task
There will be at least as many task attempts as there are tasks
If a task attempt fails, another will be started by the JobTracker
Speculative execution (see later) can also result in more task
attempts than completed tasks

03-28

MapReduce: The Mapper

Hadoop attempts to ensure that Mappers run on nodes which
hold their portion of the data locally, to avoid network traffic
Multiple Mappers run in parallel, each processing a portion of the
input data
The Mapper reads data in the form of key/value pairs
It outputs zero or more key/value pairs
map(in_key, in_value) ->
(inter_key, inter_value) list

03-29

MapReduce: The Mapper (contd)

The Mapper may use or completely ignore the input key
For example, a standard pattern is to read a line of a file at a time
The key is the byte offset into the file at which the line starts
The value is the contents of the line itself
Typically the key is considered irrelevant
If the Mapper writes anything out, the output must be in the form
of key/value pairs

03-30

Example Mapper: Upper Case Mapper

Turn input into upper case (pseudo-code):
let map(k, v) =
emit(k.toUpper(), v.toUpper())

('foo', 'bar') -> ('FOO', 'BAR')

('foo', 'other') -> ('FOO', 'OTHER')
('baz', 'more data') -> ('BAZ', 'MORE DATA')

03-31

Example Mapper: Explode Mapper

Output each input character separately (pseudo-code):
let map(k, v) =
foreach char c in v:
emit (k, c)

('foo', 'bar') ->

('foo', 'b'), ('foo', 'a'),

('foo', 'r')

('baz', 'other') -> ('baz', 'o'), ('baz', 't'),

('baz', 'h'), ('baz', 'e'),
('baz', 'r')

03-32

Example Mapper: Filter Mapper

Only output key/value pairs where the input value is a prime
number (pseudo-code):
let map(k, v) =
if (isPrime(v)) then emit(k, v)

('foo', 7) ->

('foo', 7)

('baz', 10) ->

nothing

03-33

Example Mapper: Changing Keyspaces

The key output by the Mapper does not need to be identical to
the input key
Output the word length as the key (pseudo-code):
let map(k, v) =
emit(v.length(), v)

('foo', 'bar') ->

(3, 'bar')

('baz', 'other') -> (5, 'other')

('foo', 'abracadabra') -> (11, 'abracadabra')

03-34

MapReduce: The Reducer

After the Map phase is over, all the intermediate values for a
given intermediate key are combined together into a list
This list is given to a Reducer
There may be a single Reducer, or multiple Reducers
This is specified as part of the job configuration (see later)
All values associated with a particular intermediate key are
guaranteed to go to the same Reducer
The intermediate keys, and their value lists, are passed to the
Reducer in sorted key order
This step is known as the shuffle and sort
The Reducer outputs zero or more final key/value pairs
These are written to HDFS
In practice, the Reducer usually emits a single key/value pair for
each input key
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-35

Example Reducer: Sum Reducer

Add up all the values associated with each intermediate key
(pseudo-code):
let reduce(k, vals) =
sum = 0
foreach int i in vals:
sum += i
emit(k, sum)

(bar', [9, 3, -17, 44]) ->

(bar', 39)

(foo', [123, 100, 77]) -> (foo', 300)

03-36

Example Reducer: Identity Reducer

The Identity Reducer is very common (pseudo-code):
let reduce(k, vals) =
foreach v in vals:
emit(k, v)

('foo', [9, 3, -17, 44]) ->

('foo', 9), ('foo', 3),

('foo', -17), ('foo', 44)

('bar', [123, 100, 77]) -> ('bar', 123), ('bar', 100),

('bar', 77)

03-37

MapReduce Example: Word Count

Count the number of occurrences of each word in a large amount
of input data
This is the hello world of MapReduce programming
map(String input_key, String input_value)
foreach word w in input_value:
emit(w, 1)

reduce(String output_key,
Iterator<int> intermediate_vals)
set count = 0
foreach v in intermediate_vals:
count += v
emit(output_key, count)

03-38

MapReduce Example: Word Count (contd)

Input to the Mapper:
(3414, 'the cat sat on the mat')
(3437, 'the aardvark sat on the sofa')

Output from the Mapper:

('the', 1), ('cat', 1), ('sat', 1), ('on', 1),
('the', 1), ('mat', 1), ('the', 1), ('aardvark', 1),
('sat', 1), ('on', 1), ('the', 1), ('sofa', 1)

03-39

MapReduce Example: Word Count (contd)

Intermediate data sent to the Reducer:
('aardvark', [1])
('cat', [1])
('mat', [1])
('on', [1, 1])
('sat', [1, 1])
('sofa', [1])
('the', [1, 1, 1, 1])

Final Reducer output:

('aardvark', 1)
('cat', 1)
('mat', 1)
('on', 2)
('sat', 2)
('sofa', 1)
('the', 4)

03-40

MapReduce: Data Localization

Whenever possible, Hadoop will attempt to ensure that a Map
task on a node is working on a block of data stored locally on
that node via HDFS
If this is not possible, the Map task will have to transfer the data
across the network as it processes that data
Once the Map tasks have finished, data is then transferred
across the network to the Reducers
Although the Reducers may run on the same physical machines
as the Map tasks, there is no concept of data locality for the
Reducers
All Mappers will, in general, have to communicate with all
Reducers

03-41

MapReduce: Is Shuffle and Sort a Bottleneck?

It appears that the shuffle and sort phase is a bottleneck
The reduce method in the Reducers cannot start until all
Mappers have finished
In practice, Hadoop will start to transfer data from Mappers to
Reducers as the Mappers finish work
This mitigates against a huge amount of data transfer starting as
soon as the last Mapper finishes

03-42

MapReduce: Is a Slow Mapper a Bottleneck?

It is possible for one Map task to run more slowly than the others
Perhaps due to faulty hardware, or just a very slow machine
It would appear that this would create a bottleneck
The reduce method in the Reducer cannot start until every
Mapper has finished
Hadoop uses speculative execution to mitigate against this
If a Mapper appears to be running significantly more slowly than
the others, a new instance of the Mapper will be started on
another machine, operating on the same data
The results of the first Mapper to finish will be used
Hadoop will kill off the Mapper which is still running

03-43

Hadoop: Basic Concepts

03-44

Hands-On Exercise: Running A MapReduce Job

In this Hands-On Exercise, you will run a MapReduce job on your
pseudo-distributed Hadoop cluster
Please refer to the Hands-On Exercise Manual

03-45

Hadoop: Basic Concepts

03-46

Installing A Hadoop Cluster

Cluster installation is usually performed by the system
administrator, and is outside the scope of this course
Cloudera offers a training course for System Administrators
specifically aimed at those responsible for commissioning and
maintaining Hadoop clusters
However, its very useful to understand how the component parts
of the Hadoop cluster work together
Typically, a developer will configure their machine to run in
pseudo-distributed mode
This effectively creates a single-machine cluster
All five Hadoop daemons are running on the same machine
Very useful for testing code before it is deployed to the real
cluster
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-47

Installing A Hadoop Cluster (contd)

Easiest way to download and install Hadoop, either for a full
cluster or in pseudo-distributed mode, is by using Clouderas
Distribution including Apache Hadoop (CDH)
Vanilla Hadoop plus many patches, backports, bugfixes
Supplied as a Debian package (for Linux distributions such as
Ubuntu), an RPM (for CentOS/RedHat Enterprise Linux), and as
a tarball
Full documentation available at http://cloudera.com/

03-48

The Five Hadoop Daemons

Hadoop is comprised of five separate daemons
NameNode
Holds the metadata for HDFS
Secondary NameNode
Performs housekeeping functions for the NameNode
Is not a backup or hot standby for the NameNode!
DataNode
Stores actual HDFS data blocks
JobTracker
Manages MapReduce jobs, distributes individual tasks to
machines running the
TaskTracker
Instantiates and monitors individual Map and Reduce tasks
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-49

The Five Hadoop Daemons (contd)

Each daemon runs in its own Java Virtual Machine (JVM)
No node on a real cluster will run all five daemons
Although this is technically possible
We can consider nodes to be in two different categories:
Master Nodes
Run the NameNode, Secondary NameNode, JobTracker
daemons
Only one of each of these daemons runs on the cluster
Slave Nodes
Run the DataNode and TaskTracker daemons
A slave node will run both of these daemons

03-50

Basic Cluster Configuration

03-51

Basic Cluster Configuration (contd)

On very small clusters, the NameNode, JobTracker and
Secondary NameNode can all reside on a single machine
It is typical to put them on separate machines as the cluster
grows beyond 20-30 nodes
Each dotted box on the previous diagram represents a separate
Java Virtual Machine (JVM)

03-52

Submitting A Job
When a client submits a job, its configuration information is
packaged into an XML file
This file, along with the .jar file containing the actual program
code, is handed to the JobTracker
The JobTracker then parcels out individual tasks to TaskTracker
nodes
When a TaskTracker receives a request to run a task, it
instantiates a separate JVM for that task
TaskTracker nodes can be configured to run multiple tasks at the
same time
If the node has enough processing power and memory

03-53

Submitting A Job (contd)

The intermediate data is held on the TaskTrackers local disk
As Reducers start up, the intermediate data is distributed across
the network to the Reducers
Reducers write their final output to HDFS
Once the job has completed, the TaskTracker can delete the
intermediate data from its local disk
Note that the intermediate data is not deleted until the entire job
completes

03-54

Hadoop: Basic Concepts

03-55

Other Ecosystem Projects: Introduction

The term Hadoop refers to HDFS and MapReduce
Many other projects exist which use Hadoop
Either both HDFS and MapReduce, or just HDFS
Many are Apache projects or Apache Incubator projects
Some others are not hosted by the Apache Software Foundation
These are often hosted on GitHub or a similar repository
We will investigate many of these projects later in the course
Following is an introduction to some of the most significant
projects

03-56

Hive
Hive is an abstraction on top of MapReduce
Allows users to query data in the Hadoop cluster without
knowing Java or MapReduce
Uses the HiveQL language
Very similar to SQL
The Hive Interpreter runs on a client machine
Turns HiveQL queries into MapReduce jobs
Submits those jobs to the cluster
Note: this does not turn the cluster into a relational database
server!
It is still simply running MapReduce jobs
Those jobs are created by the Hive Interpreter
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-57

Hive (contd)
Sample Hive query:
SELECT stock.product, SUM(orders.purchases)
FROM stock INNER JOIN orders
ON (stock.id = orders.stock_id)
WHERE orders.quarter = 'Q1'
GROUP BY stock.product;
We will investigate Hive in greater detail later in the course

03-58

Pig
Pig is an alternative abstraction on top of MapReduce
Uses a dataflow scripting language
Called PigLatin
The Pig interpreter runs on the client machine
Takes the PigLatin script and turns it into a series of MapReduce
jobs
Submits those jobs to the cluster
As with Hive, nothing magical happens on the cluster
It is still simply running MapReduce jobs

03-59

Pig (contd)
Sample Pig script:
stock = LOAD '/user/fred/stock' AS (id, item);
orders= LOAD '/user/fred/orders' AS (id, cost);
grpd = GROUP orders BY id;
totals = FOREACH grpd GENERATE group,
SUM(orders.cost) AS t;
result = JOIN stock BY id, totals BY group;
DUMP result;
We will investigate Pig in more detail later in the course

03-60

Flume and Sqoop

Flume provides a method to import data into HDFS as it is
generated
Rather than batch-processing the data later
For example, log files from a Web server
Sqoop provides a method to import data from tables in a
relational database into HDFS
Does this very efficiently via a Map-only MapReduce job
Can also go the other way
Populate database tables from files in HDFS
We will investigate Flume and Sqoop later in the course

03-61

Oozie
Oozie allows developers to create a workflow of MapReduce jobs
Including dependencies between jobs
The Oozie server submits the jobs to the server in the correct
sequence
We will investigate Oozie later in the course

03-62

HBase
HBase is the Hadoop database
A NoSQL datastore
Can store massive amounts of data
Gigabytes, terabytes, and even petabytes of data in a table
Scales to provide very high write throughput
Hundreds of thousands of inserts per second
Copes well with sparse data
Tables can have many thousands of columns
Even if most columns are empty for any given row
Has a very constrained access model
Insert a row, retrieve a row, do a full or partial table scan
Only one column (the row key) is indexed
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

03-63

HBase vs Traditional RDBMSs

RDBMS&

HBase&

Data&layout&

Row1oriented&

Column1oriented&

Transac:ons&

Yes&

Single&row&only&

Query&language&

SQL&

get/put/scan&

Security&

Authen:ca:on/Authoriza:on&

TBD&

Indexes&

On&arbitrary&columns&

Row1key&only&

Max&data&size&

TBs&

PB+&

Read/write&throughput&
limits&

1000s&queries/second&

Millions&of&queries/second&

03-64

Hadoop: Basic Concepts

03-65

Conclusion
In this chapter you have learned
What Hadoop is
What features the Hadoop Distributed File System (HDFS)
provides
The concepts behind MapReduce
How a Hadoop cluster operates
What other Hadoop Ecosystem projects exist

03-66

Chapter 4
Writing a MapReduce
Program

04-1

04-2

Writing a MapReduce Program

In this chapter you will learn
How to use the Hadoop API to write a MapReduce program in
Java
How to use the Streaming API to write Mappers and Reducers in
other languages
How to use Eclipse to speed up your Hadoop development
The differences between the Old and New Hadoop APIs

04-3

Writing a MapReduce Program

The MapReduce Flow
Examining our Sample MapReduce program
The Driver Code
The Mapper
The Reducer
Hadoops Streaming API
Using Eclipse for Rapid Development
Hands-On Exercise: Write a MapReduce program
The New MapReduce API
Conclusion
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-4

A Sample MapReduce Program: Introduction

In the previous chapter, you ran a sample MapReduce program
WordCount, which counted the number of occurrences of each
unique word in a set of files
In this chapter, we will examine the code for WordCount
This will demonstrate the Hadoop API
We will also investigate Hadoop Streaming
Allows you to write MapReduce programs in (virtually) any
language

04-5

The MapReduce Flow: Introduction

On the following slides we show the MapReduce flow
Each of the portions (RecordReader, Mapper, Partitioner,
Reducer, etc.) can be created by the developer
We will cover each of these as we move through the course
You will always create at least a Mapper, Reducer, and driver
code
Those are the portions we will investigate in this chapter

04-6

The MapReduce Flow: The Mapper

04-7

The MapReduce Flow: Shuffle and Sort

04-8

The MapReduce Flow: Reducers to Outputs

04-9

Writing a MapReduce Program

04-10

Our MapReduce Program: WordCount

To investigate the API, we will dissect the WordCount program
you ran in the previous chapter
This consists of three portions
The driver code
Code that runs on the client to configure and submit the job
The Mapper
The Reducer
Before we look at the code, we need to cover some basic Hadoop
API concepts

04-11

Getting Data to the Mapper

The data passed to the Mapper is specified by an InputFormat
Specified in the driver code
Defines the location of the input data
A file or directory, for example
Determines how to split the input data into input splits
Each Mapper deals with a single input split
InputFormat is a factory for RecordReader objects to extract
(key, value) records from the input source

04-12

Getting Data to the Mapper (contd)

04-13

Some Standard InputFormats

FileInputFormat
The base class used for all file-based InputFormats
TextInputFormat
The default
Treats each \n-terminated line of a file as a value
Key is the byte offset within the file of that line
KeyValueTextInputFormat
Maps \n-terminated lines as key SEP value
By default, separator is a tab
SequenceFileInputFormat
Binary file of (key, value) pairs with some additional metadata
SequenceFileAsTextInputFormat
Similar, but maps (key.toString(), value.toString())
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-14

Keys and Values are Objects

Keys and values in Hadoop are Objects
Values are objects which implement Writable
Keys are objects which implement WritableComparable

04-15

What is Writable?
Hadoop defines its own box classes for strings, integers and so
on
IntWritable for ints
LongWritable for longs
FloatWritable for floats
DoubleWritable for doubles
Text for strings
Etc.
The Writable interface makes serialization quick and easy for
Hadoop
Any values type must implement the Writable interface

04-16

What is WritableComparable?
A WritableComparable is a Writable which is also
Comparable
Two WritableComparables can be compared against each
other to determine their order
Keys must be WritableComparables because they are passed
to the Reducer in sorted order
We will talk more about WritableComparable later
Note that despite their names, all Hadoop box classes implement
both Writable and WritableComparable
For example, IntWritable is actually a
WritableComparable

04-17

Writing a MapReduce Program

04-18

The Driver Code: Introduction

The driver code runs on the client machine
It configures the job, then submits it to the cluster

04-19

The Driver: Complete Code

import
import
import
import
import
import
import
import
import
import

org.apache.hadoop.fs.Path;
org.apache.hadoop.io.IntWritable;
org.apache.hadoop.io.Text;
org.apache.hadoop.mapred.FileInputFormat;
org.apache.hadoop.mapred.FileOutputFormat;
org.apache.hadoop.mapred.JobClient;
org.apache.hadoop.mapred.JobConf;
org.apache.hadoop.conf.Configured;
org.apache.hadoop.util.Tool;
org.apache.hadoop.util.ToolRunner;

public class WordCount extends Configured implements Tool {

public int run(String[] args) throws Exception {
if (args.length != 2) {
System.out.printf(
"Usage: %s [generic options] <input dir> <output dir>\n",
getClass().getSimpleName());
ToolRunner.printGenericCommandUsage(System.out);
return -1;
}
JobConf conf = new JobConf(getConf(), WordCount.class);
conf.setJobName(this.getClass().getName());
FileInputFormat.setInputPaths(conf, new Path(args[0]));
FileOutputFormat.setOutputPath(conf, new Path(args[1]));
conf.setMapperClass(WordMapper.class);
conf.setReducerClass(SumReducer.class);
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-20

The Driver: Complete Code (contd)

conf.setMapOutputKeyClass(Text.class);
conf.setMapOutputValueClass(IntWritable.class);
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}

04-21

The Driver: Import Statements

import
import
import
import
import
import
import
import
import
import

public class WordCount extends Configured implements Tool {

public int run(String[] args) throws Exception {

You will typically import these classes into every

MapReduce job you write. We will omit the import
statements in future slides for brevity.

if (args.length != 2) {
System.out.printf(
"Usage: %s [generic options] <input dir> <output dir>\n",
getClass().getSimpleName());
ToolRunner.printGenericCommandUsage(System.out);
return -1;
}
JobConf conf = new JobConf(getConf(), WordCount.class);
conf.setJobName(this.getClass().getName());
FileInputFormat.setInputPaths(conf, new Path(args[0]));
FileOutputFormat.setOutputPath(conf, new Path(args[1]));
conf.setMapperClass(WordMapper.class);
conf.setReducerClass(SumReducer.class);

04-22

The Driver: Main Code

public class WordCount extends Configured implements Tool {
public int run(String[] args) throws Exception {
if (args.length != 2) {
System.out.printf(
"Usage: %s [generic options] <input dir> <output dir>\n", getClass().getSimpleName());
ToolRunner.printGenericCommandUsage(System.out);
return -1;
}
JobConf conf = new JobConf(getConf(), WordCount.class);
conf.setJobName(this.getClass().getName());
FileInputFormat.setInputPaths(conf, new Path(args[0]));
FileOutputFormat.setOutputPath(conf, new Path(args[1]));
conf.setMapperClass(WordMapper.class);
conf.setReducerClass(SumReducer.class);
conf.setMapOutputKeyClass(Text.class);
conf.setMapOutputValueClass(IntWritable.class);
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-23

The Driver Class: Using ToolRunner

Your driver class extends Configured and implements Tool.

This allows the user to specify configuration settings on the
command line, which will then be incorporated into the jobs
configuration when it is submitted to the server. Although this
FileInputFormat.setInputPaths(conf, new Path(args[0]));
FileOutputFormat.setOutputPath(conf, new Path(args[1]));
is not compulsory, it is considered a best practice. (We will
conf.setMapperClass(WordMapper.class);
discuss ToolRunner in more detail later.)
conf.setReducerClass(SumReducer.class);
conf.setMapOutputKeyClass(Text.class);
conf.setMapOutputValueClass(IntWritable.class);
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-24

The Driver Class: Using ToolRunner (contd)

The main method simply calls ToolRunner.run(), passing

conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);
in
the driver class and the command-line arguments. The job
JobClient.runJob(conf);
will
then be configured and submitted in the run method.
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-25

Sanity Checking The Jobs Invocation

public class WordCount extends Configured implements Tool {
public int run(String[] args) throws Exception {
if (args.length != 2) {
System.out.printf(
"Usage: %s [generic options] <input dir> <output dir>\n", getClass().getSimpleName());
ToolRunner.printGenericCommandUsage(System.out);
return -1;
}
JobConf conf = new JobConf(getConf(), WordCount.class);
conf.setJobName(this.getClass().getName());
FileInputFormat.setInputPaths(conf,
new Path(args[0]));
The
first step is to ensure that
we have been given two
FileOutputFormat.setOutputPath(conf, new Path(args[1]));
command-line
arguments. If not, print a help message and
conf.setMapperClass(WordMapper.class);
conf.setReducerClass(SumReducer.class);
exit.
conf.setMapOutputKeyClass(Text.class);
conf.setMapOutputValueClass(IntWritable.class);
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-26

Configuring The Job With JobConf

FileOutputFormat.setOutputPath(conf,
Path(args[1]));
To
configure the job, create new
a new
JobConf object and specify
conf.setMapperClass(WordMapper.class);
the
class which will be called to run the job.
conf.setReducerClass(SumReducer.class);
conf.setMapOutputKeyClass(Text.class);
conf.setMapOutputValueClass(IntWritable.class);
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-27

Creating a New JobConf Object

The JobConf class allows you to set configuration options for
your MapReduce job
The classes to be used for your Mapper and Reducer
The input and output directories
Many other options
Any options not explicitly set in your driver code will be read
from your Hadoop configuration files
Usually located in /etc/hadoop/conf
Any options not specified in your configuration files will receive
Hadoops default values

04-28

Naming The Job

Give the job a meaningful name.

conf.setMapperClass(WordMapper.class);
conf.setReducerClass(SumReducer.class);
conf.setMapOutputKeyClass(Text.class);
conf.setMapOutputValueClass(IntWritable.class);
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-29

Specifying Input and Output Directories

Next, specify the input directory from which data will be read,
and
the output directory to which final output will be written.
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-30

Specifying the InputFormat

The default InputFormat (TextInputFormat) will be used unless
you specify otherwise
To use an InputFormat other than the default, use e.g.
conf.setInputFormat(KeyValueTextInputFormat.class)

04-31

Determining Which Files To Read

By default, FileInputFormat.setInputPaths() will read all
files from a specified directory and send them to Mappers
Exceptions: items whose names begin with a period (.) or
underscore (_)
Globs can be specified to restrict input
For example, /2010/*/01/*
Alternatively, FileInputFormat.addInputPath() can be called
multiple times, specifying a single file or directory each time
More advanced filtering can be performed by implementing a
PathFilter
Interface with a method named accept
Takes a path to a file, returns true or false depending on
whether or not the file should be processed
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-32

Specifying Final Output With OutputFormat

FileOutputFormat.setOutputPath() specifies the directory
to which the Reducers will write their final output
The driver can also specify the format of the output data
Default is a plain text file
Could be explicitly written as
conf.setOutputFormat(TextOutputFormat.class);
We will discuss OutputFormats in more depth in a later chapter

04-33

Specify The Classes for Mapper and Reducer

Give
the JobConf object information about which classes are
conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);
to
be instantiated as the Mapper and Reducer.
JobClient.runJob(conf);
}

return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-34

Specify The Intermediate Data Types

Specify the types of the intermediate output key and value

JobClient.runJob(conf);
return 0;
produced
by the Mapper.
}

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-35

Specify The Final Output Data Types

JobClient.runJob(conf);
return 0;

Specify the types of the Reducers output key and value.

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-36

Running The Job

Finally, run the job by calling the runJob method.

conf.setOutputKeyClass(Text.class);
conf.setOutputValueClass(IntWritable.class);

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-37

Running The Job (contd)

There are two ways to run your MapReduce job:
JobClient.runJob(conf)
Blocks (waits for the job to complete before continuing)
JobClient.submitJob(conf)
Does not block (driver code continues as the job is running)
JobClient determines the proper division of input data into
InputSplits
JobClient then sends the job information to the JobTracker
daemon on the cluster

04-38

Reprise: Driver Code

JobClient.runJob(conf);
return 0;

public static void main(String[] args) throws Exception {

int exitCode = ToolRunner.run(new WordCount(), args);
System.exit(exitCode);
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-39

Writing a MapReduce Program

04-40

The Mapper: Complete Code

import java.io.IOException;
import
import
import
import
import
import
import

org.apache.hadoop.io.IntWritable;
org.apache.hadoop.io.LongWritable;
org.apache.hadoop.io.Text;
org.apache.hadoop.mapred.MapReduceBase;
org.apache.hadoop.mapred.Mapper;
org.apache.hadoop.mapred.OutputCollector;
org.apache.hadoop.mapred.Reporter;

public class WordMapper extends MapReduceBase implements

Mapper<LongWritable, Text, Text, IntWritable> {

public void map(LongWritable key, Text value,

OutputCollector<Text, IntWritable> output, Reporter reporter)
throws IOException {
String s = value.toString();
for (String word : s.split("\\W+")) {
if (word.length() > 0) {
output.collect(new Text(word), new IntWritable(1));
}
}
}

04-41

The Mapper: import Statements

import java.io.IOException;
import
import
import
import
import
import
import

public class WordMapper extends MapReduceBase implements

Mapper<LongWritable, Text, Text, IntWritable> {

You will typically import java.io.IOException, and

public void
map(LongWritable
key, Text classes
value, shown, in every
the
org.apache.hadoop
OutputCollector<Text, IntWritable> output, Reporter reporter)
throwsMapper
IOException
you {write. We will omit the import statements
String s = value.toString();
for (String
word :
s.split("\\W+"))
in future
slides
for brevity. {
if (word.length() > 0) {
output.collect(new Text(word), new IntWritable(1));
}

04-42

The Mapper: Main Code

public class WordMapper extends MapReduceBase
implements Mapper<LongWritable, Text, Text, IntWritable> {

public void map(LongWritable key, Text value,

04-43

The Mapper: Main Code (contd)

public class WordMapper extends MapReduceBase
implements Mapper<LongWritable, Text, Text, IntWritable> {

public void map(LongWritable key, Text value,

Your Mapper class should
extend MapReduceBase,
OutputCollector<Text,
IntWritable>
output, Reporter reporter)
throws IOException {
and implement the Mapper interface. The Mapper
String s = value.toString();
for (String
: s.split("\\W+"))
{
interfaceword
expects
four generics, which
define the types
if (word.length() > 0) {
of the input and output
key/value pairs.
The first two
output.collect(new
Text(word),
new IntWritable(1));
} parameters define the input key and value types, the
}
second two define the output key and value types.
}

04-44

The map Method

public class WordMapper extends MapReduceBase
implements Mapper<LongWritable, Text, Text, IntWritable> {

public void map(LongWritable key, Text value,

OutputCollector<Text, IntWritable> output, Reporter reporter)
throws IOException {
String s = value.toString();
for (String word : s.split("\\W+")) {
ifThe
(word.length()
0) {
map methods> signature
looks like this. It will be
output.collect(new Text(word), new IntWritable(1));
passed
a key, a value, an OutputCollector object
}
}
and a Reporter object. The OutputCollector is
}

used to write the intermediate data; you must specify

the data types that it will write.

04-45

The map Method: Processing The Line

public class WordMapper extends MapReduceBase
implements Mapper<LongWritable, Text, Text, IntWritable> {

public void map(LongWritable key, Text value,

OutputCollector<Text, IntWritable> output, Reporter reporter)
throws IOException {
String s = value.toString();
for (String word : s.split("\\W+")) {
if (word.length() > 0) {
value
is a Text object, so
we retrievenew
theIntWritable(1));
string it
output.collect(new
Text(word),
}
contains.
}
}

04-46

The map Method: Processing The Line (contd)

public class WordMapper extends MapReduceBase
implements Mapper<LongWritable, Text, Text, IntWritable> {

public void map(LongWritable key, Text value,

alphanumeric characters as the word delimiter, and

loop through those words.

04-47

Outputting Intermediate Data

public class WordMapper extends MapReduceBase
implements Mapper<LongWritable, Text, Text, IntWritable> {

public void map(LongWritable key, Text value,

OutputCollector object. The key will be the word itself, the

value will be the number 1. Recall that the output key must be of
type WritableComparable, and the value must be a Writable.

04-48

Reprise: The Map Method

public class WordMapper extends MapReduceBase
implements Mapper<LongWritable, Text, Text, IntWritable> {

public void map(LongWritable key, Text value,

04-49

The Reporter Object

Notice that in this example we have not used the Reporter
object which was passed to the Mapper
The Reporter object can be used to pass some information back
to the driver code
We will investigate the Reporter later in the course

04-50

Writing a MapReduce Program

04-51

The Reducer: Complete Code

import java.io.IOException;
import java.util.Iterator;
import
import
import
import
import
import

org.apache.hadoop.io.IntWritable;
org.apache.hadoop.io.Text;
org.apache.hadoop.mapred.OutputCollector;
org.apache.hadoop.mapred.MapReduceBase;
org.apache.hadoop.mapred.Reducer;
org.apache.hadoop.mapred.Reporter;

public class SumReducer extends MapReduceBase implements

Reducer<Text, IntWritable, Text, IntWritable> {
public void reduce(Text key, Iterator<IntWritable> values,
OutputCollector<Text, IntWritable> output, Reporter reporter)
throws IOException {

int wordCount = 0;
while (values.hasNext()) {
IntWritable value = values.next();
wordCount += value.get();
}
output.collect(key, new IntWritable(wordCount));

04-52

The Reducer: Import Statements

import java.io.IOException;
import java.util.Iterator;
import
import
import
import
import
import

public class SumReducer extends MapReduceBase implements

Reducer<Text, IntWritable, Text, IntWritable> {

As with the Mapper, you will typically import

public void reduce(Text key, Iterator<IntWritable> values,
java.io.IOException,
and the
org.apache.hadoop
OutputCollector<Text,
IntWritable> output,
Reporter
reporter)
throws IOException {
classes shown, in every Reducer you write. You will also
int wordCount = 0;
import java.util.Iterator,
which will be used to step
while (values.hasNext())
{
IntWritable value = values.next();
through
the values provided to the Reducer for each key.
wordCount
+= value.get();
}
We will omit
import statements in future slides for
output.collect(key,
newthe
IntWritable(wordCount));
}
brevity.
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-53

The Reducer: Main Code

public class SumReducer extends MapReduceBase implements
Reducer<Text, IntWritable, Text, IntWritable> {
public void reduce(Text key, Iterator<IntWritable> values,
OutputCollector<Text, IntWritable> output, Reporter reporter)
throws IOException {

int wordCount = 0;
while (values.hasNext()) {
IntWritable value = values.next();
wordCount += value.get();
}
output.collect(key, new IntWritable(wordCount));

04-54

The Reducer: Main Code (contd)

public class SumReducer extends MapReduceBase implements
Reducer<Text, IntWritable, Text, IntWritable> {
public void reduce(Text key, Iterator<IntWritable> values,
Your Reducer class should
extend MapReduceBase
OutputCollector<Text,
IntWritable>
output, Reporter reporter)
throws IOException {

and implement Reducer. The Reducer interface

int wordCount
= generics,
0;
expects four
which define the types of the
while (values.hasNext()) {
input and output
pairs. The first two
IntWritable
value key/value
= values.next();
wordCount
+= value.get();
parameters
define the intermediate key and value
}
types, the secondnew
twoIntWritable(wordCount));
define the final output key and
output.collect(key,
value types. The keys are WritableComparables,
the values are Writables.

04-55

The reduce Method

int wordCount = 0;
while (values.hasNext())
{
The reduce method
receives a key and an Iterator of
IntWritable value = values.next();
it also receives an OutputCollector object
wordCountvalues;
+= value.get();
}
and a Reporter object.
output.collect(key, new IntWritable(wordCount));

04-56

Processing The Values

int wordCount = 0;
while (values.hasNext()) {
IntWritable value = values.next();
wordCount += value.get();
}
output.collect(key, new IntWritable(wordCount));

We use the hasNext() and next() methods on

values to step through all the elements in the iterator.
In our example, we are merely adding all the values
together. We use value().get() to retrieve the
actual numeric value.
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-57

Writing The Final Output

int wordCount = 0;
while (values.hasNext()) {
IntWritable value = values.next();
wordCount += value.get();
}
output.collect(key, new IntWritable(wordCount));

Finally, we write the output (key, value) pair using the

collect method of our OutputCollector object.

04-58

Reprise: The Reduce Method

int wordCount = 0;
while (values.hasNext()) {
IntWritable value = values.next();
wordCount += value.get();
}
output.collect(key, new IntWritable(wordCount));

04-59

Writing a MapReduce Program

04-60

The Streaming API: Motivation

Many organizations have developers skilled in languages other
than Java, such as
Ruby
Python
Perl
The Streaming API allows developers to use any language they
wish to write Mappers and Reducers
As long as the language can read from standard input and write
to standard output

04-61

The Streaming API: Advantages

Advantages of the Streaming API:
No need for non-Java coders to learn Java
Fast development time
Ability to use existing code libraries

04-62

How Streaming Works

To implement streaming, write separate Mapper and Reducer
programs in the language of your choice
They will receive input via stdin
They should write their output to stdout
If TextInputFormat (the default) is used, the streaming Mapper
just receives each line from the file on stdin
No key is passed
Streaming Mapper and streaming Reducers output should be
sent to stdout as key (tab) value (newline)
Separators other than tab can be specified

04-63

Streaming: Example Mapper

Example streaming wordcount Mapper:
#!/usr/bin/env perl
while (<>) {
chomp;
(@words) = split /\s+/;
foreach $w (@words) {
print "$w\t1\n";
}
}

04-64

Streaming Reducers: Caution

Recall that in Java, all the values associated with a key are
passed to the Reducer as an Iterator
Using Hadoop Streaming, the Reducer receives its input as (key,
value) pairs
One per line of standard input
Your code will have to keep track of the key so that it can detect
when values from a new key start appearing

04-65

Launching a Streaming Job

To launch a Streaming job, use e.g.,:
hadoop jar $HADOOP_HOME/contrib/streaming/hadoop-streaming*.jar \
-input myInputDirs \
-output myOutputDir \
-mapper myMapScript.pl \
-reducer myReduceScript.pl \
-file myMapScript.pl \
-file myReduceScript.pl

Many other command-line options are available

See the documentation for full details
Note that system commands can be used as a Streaming Mapper
or Reducer
For example: awk, grep, sed, or wc
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

04-66

Writing a MapReduce Program

04-67

Integrated Development Environments

There are many Integrated Development Environments (IDEs)
available
Eclipse is one such IDE
Open source
Very popular among Java developers
Has plug-ins to speed development in several different languages
If you would prefer to write your code this week using a terminalbased editor such as vi, we certainly wont stop you!
But using Eclipse can dramatically speed up your development
process
On the next few slides we will demonstrate how to use Eclipse to
write a MapReduce program

04-68

Using Eclipse
Launch Eclipse by double-clicking on the Eclipse icon on the
desktop
If you are asked whether you want to send usage data, hit
Cancel

04-69

Using Eclipse (contd)

Expand the Hadoop project

Right-click on src, and choose New Class

04-70

Using Eclipse (contd)

Enter a class name, check
the box to generate the
main method stub, and
then click Finish

04-71

Using Eclipse (contd)

You can now edit your class

04-72

Using Eclipse (contd)

Add other classes in the same way. When you are ready to test
your code, right-click on the default package and choose Export

04-73

Using Eclipse (contd)

Expand Java, select
the JAR file item, and
then click Next.

04-74

Using Eclipse (contd)

Enter a path and filename
inside /home/training
(your home directory), and
click Finish
Your JAR file will be saved; you
can now run it from the command
line with the standard
hadoop jar... command

04-75

Writing a MapReduce Program

04-76

Hands-On Exercise: Write A MapReduce

Program
In this Hands-On Exercise, you will write a MapReduce program
using either Java or Hadoops Streaming interface
Please refer to the Hands-On Exercise Instructions

04-77

Writing a MapReduce Program

04-78

What Is The New API?

When Hadoop 0.20 was released, a New API was introduced
Designed to make the API easier to evolve in the future
Favors abstract classes over interfaces
The Old API was deprecated
However, the New API is still not absolutely feature-complete in
Hadoop 0.20.x
The Old API should not have been deprecated as quickly as it
was
Most developers still use the Old API
All the code examples in this course use the Old API

04-79

Old API vs New API: Some Key Differences

Old API

New API

import org.apache.hadoop.mapred.*

import org.apache.hadoop.mapreduce.*

Driver code:

JobConf conf = new JobConf(conf,

Driver.class);
conf.setSomeProperty(...);
...
JobClient.runJob(conf);

Configuration conf = new Configuration;

Job job = new Job(conf);
job.setJarByClass(Driver.class);
job.setSomeProperty(...);
...
job.waitForCompletion(true);

Mapper:

public class MyMapper extends MapReduceBase

implements Mapper {

public class MyMapper extends Mapper {

public void map(Keytype k, Valuetype v,
Context c) {
...
c.write(key, val);
}

public void map(Keytype k, Valuetype v,

OutputCollector o, Reporter r) {
...
o.collect(key, val);
}
}
}

04-80

Old API vs New API: Some Key Differences

(contd)
Old API

New API

Reducer:

public class MyReducer extends MapReduceBase

implements Reducer {

public class MyReducer extends Reducer {

public void reduce(Keytype k,

Iterator<Valuetype> v,
OutputCollector o, Reporter r) {
while(v.hasnext()) {
// process v.next()
o.collect(key, val);
}
}

public void reduce(Keytype k,

Iterable<Valuetype> v, Context c) {
for(Valuetype v : eachval) {
// process eachval
c.write(key, val);
}
}
}

}
configure(JobConf job) (See later)

setup(Context c)

close() (See later)

cleanup(Context c)

04-81

Writing a MapReduce Program

04-82

Conclusion
In this chapter you have learned
How to use the Hadoop API to write a MapReduce program in
Java
How to use the Streaming API to write Mappers and Reducers in
other languages
How to use Eclipse to speed up your Hadoop development
The differences between the Old and New Hadoop APIs

04-83

Chapter 5
Integrating Hadoop Into
The Workflow

05-1

05-2

Integrating Hadoop Into The Workflow

In this chapter you will learn
How Hadoop can be integrating into an existing enterprise
How to load data from an existing RDBMS into HDFS using
Sqoop
How to manage real-time data such as log files using Flume
How to access HDFS from legacy systems with FuseDFS and
Hoop

05-3

Integrating Hadoop Into The Workflow

Introduction
Relational Database Management Systems
Storage Systems
Importing Data From RDBMSs With Sqoop
Hands-On Exercise
Importing Real-Time Data With Flume
Accessing HDFS Using FuseDFS and Hoop
Conclusion
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

05-4

Introduction
Your data center already has a lot of components
Database servers
Data warehouses
File servers
Backup systems
How does Hadoop fit into this ecosystem?

05-5

Integrating Hadoop Into The Workflow

05-6

RDBMS Strengths
Relational Database Management Systems (RDBMSs) have many
strengths
Ability to handle complex transactions
Ability to process hundreds or thousands of queries per second
Real-time delivery of results
Simple but powerful query language

05-7

RDBMS Weaknesses
There are some areas where RDBMSs are less ideal
Data schema is determined before data is ingested
Can make ad-hoc data collection difficult
Upper bound on data storage of 100s of terabytes
Practical upper bound on data in a single query of 10s of
terabytes

05-8

Typical RDBMS Scenario

Typical scenario: use an interactive RDBMS to serve queries
from a Web site etc
Data is later extracted and loaded into a data warehouse for
future processing and archiving
Usually denormalized into an OLAP cube

OLAP: OnLine Analytical Processing

05-9

Typical RDBMS Scenario (contd)

05-10

OLAP Database Limitations

All dimensions must be prematerialized
Re-materialization can be very time consuming
Daily data load-in times can increase
Typically this leads to some data being discarded

05-11

Using Hadoop to Augment Existing Databases

05-12

Benefits of Hadoop
Processing power scales with data storage
As you add more nodes for storage, you get more processing
power for free
Views do not need prematerialization
Ad-hoc full or partial dataset queries are possible
Total query size can be multiple petabytes

05-13

Hadoop Tradeoffs
Cannot serve interactive queries
The fastest Hadoop job will still take several seconds to run
Less powerful updates
No transactions
No modification of existing records

05-14

Integrating Hadoop Into The Workflow

05-15

Traditional High-Performance File Servers

Enterprise data is often held on large fileservers, such as
NetApp
EMC
Advantages:
Fast random access
Many concurrent clients
Disadvantages
High cost per terabyte of storage

05-16

File Servers and Hadoop

Choice of destination medium depends on the expected access
patterns
Sequentially read, append-only data: HDFS
Random access: file server
HDFS can crunch sequential data faster
Offloading data to HDFS leaves more room on file servers for
interactive data
Use the right tool for the job!

05-17

Integrating Hadoop Into The Workflow

05-18

Importing Data From an RDBMS to HDFS

Typical scenario: the need to use data stored in an RDBMS (such
as Oracle database, MySQL or Teradata) in a MapReduce job
Lookup tables
Legacy data
Possible to read directly from an RDBMS in your Mapper
Can lead to the equivalent of a distributed denial of service
(DDoS) attack on your RDBMS
In practice dont do it!
Better scenario: import the data into HDFS beforehand

05-19

Sqoop: SQL to Hadoop

Sqoop: open source tool written at Cloudera
Imports tables from an RDBMS into HDFS
Just one table
All tables in a database
Just portions of a table
Sqoop supports a WHERE clause
Uses MapReduce to actually import the data
Throttles the number of Mappers to avoid DDoS scenarios
Uses four Mappers by default
Value is configurable
Uses a JDBC interface
Should work with any JDBC-compatible database
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

05-20

Sqoop: SQL to Hadoop (contd)

Imports data to HDFS as delimited text files or SequenceFiles
Default is a comma-delimited text file
Can be used for incremental data imports
First import retrieves all rows in a table
Subsequent imports retrieve just rows created since the last
import
Generates a class file which can encapsulate a row of the
imported data
Useful for serializing and deserializing data in subsequent
MapReduce jobs

05-21

Custom Sqoop Connectors

Cloudera has partnered with other organizations to create
custom Sqoop connectors
Use a systems native protocols to access data rather than JDBC
Provides much faster performance
Current systems supported by custom connectors include:
Netezza
Teradata
MicroStrategy
Oracle Database (connector developed with Quest Software)
Others are in development
Custom connectors are not open source, but are free
Available from the Cloudera Web site
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

05-22

Sqoop: Basic Syntax

Standard syntax:
sqoop tool-name [tool-options]

Tools include:
import
import-all-tables
list-tables

Options include:
--connect
--username
--password
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

05-23

Sqoop: Example
Example: import a table called employees from a database called
personnel in a MySQL RDBMS
sqoop import --username fred --password derf \
--connect jdbc:mysql://database.example.com/personnel
--table employees

Example: as above, but only records with an id greater than 1000

sqoop import --username fred --password derf \
--connect jdbc:mysql://database.example.com/personnel
--table employees
--where "id > 1000"

05-24

Sqoop: Importing To Hive Tables

The Sqoop option --hive-import will automatically create a
Hive table from the imported data
Imports the data
Generates the Hive CREATE TABLE statement
Runs the statement
Note: This will move the imported table into Hives warehouse
directory

05-25

Sqoop: Other Options

Sqoop can take data from HDFS and insert it into an alreadyexisting table in an RDBMS with the command
sqoop export [options]

For general Sqoop help:

sqoop help

For help on a particular command:

sqoop help command

05-26

Integrating Hadoop Into The Workflow

05-27

Hands-On Exercise: Importing Data

In this Hands-On Exercise, you will import data into HDFS from
MySQL
Please refer to the Hands-On Exercise Manual

05-28

Integrating Hadoop Into The Workflow

05-29

Flume: Basics
Flume is a distributed, reliable, available service for efficiently
moving large amounts of data as it is produced
Ideally suited to gathering logs from multiple systems and
inserting them into HDFS as they are generated
Flume is Open Source
Initially developed by Cloudera
Flumes design goals:
Reliability
Scalability
Manageability
Extensibility

05-30

Flume: High-Level Overview

Agent&

Agent&&

Agent&

encrypt&

MASTER&
Master'communicates'
with'all'Agents,'
specifying'congura$on'
etc.'

Processor&

compress&

batch&
encrypt&

Writes'to'mul$ple'HDFS'le'
formats'(text,'SequenceFile,'
JSON,'Avro,'others)'
Parallelized'writes'across'many'
collectors''as'much'write'
throughput'as'required'

Collector(s)&

Mul$ple'congurable'
levels'of'reliability'
Agents''can'guarantee'
delivery'in'event'of'
failure'
Op$onally'deployable,'
centrally'administered'

Op$onally'pre=process'
incoming'data:'perform'
transforma$ons,'
suppressions,'metadata'
enrichment'
Flexibly'
deploy'
decorators'at'
any'step'to'
improve'
performance,''
reliability'or'
security'

05-31

Flume Node Characteristics

Each Flume node has a source and a sink
Source
Tells the node where to receive data from
Sink
Tells the node where to send data to
Sink can have one or more decorators
Perform simple processing on data as it passes though, such as
Compression
Encryption
awk, grep-like functionality

05-32

Flumes Design Goals: Reliability

Reliability
The ability to continue delivering events in the face of system
component failure
Provides user-configurable reliability guarantees
End-to-end
Once Flume acknowledges receipt of an event, the event will
eventually make it to the end of the pipeline
Store on failure
Nodes require acknowledgment of receipt from the node one
hop downstream
Best effort
No attempt is made to confirm receipt of data from the node
downstream
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

05-33

Flumes Design Goals: Scalability

Scalability
The ability to increase system performance linearly or better
by adding more resources to the system
Flume scales horizontally
As load increases, more machines can be added to the
configuration

05-34

Flumes Design Goals: Manageability

Manageability
The ability to control data flows, monitor nodes, modify the
settings, and control outputs of a large system
Flume provides a central Master, where users can monitor data
flows and reconfigure them on the fly
Via a Web interface or a scriptable command-line shell

05-35

Flumes Design Goals: Extensibility

Extensibility
The ability to add new functionality to a system
Flume can be extended by adding connectors to existing storage
layers or data platforms
General sources include data from files, syslog, and standard
output from a process
General endpoints include files on the local filesystem or HDFS
Other connectors can be added
IRC
Twitter streams
HBase
Developers can write their own connectors in Java

05-36

Flume: Usage Patterns

Flume is typically used to ingest log files from real-time systems
such as Web servers, firewalls and mailservers into HDFS
Currently in use in many large organizations, ingesting millions
of events per day
At least one organization is using Flume to ingest over 200 million
events per day
Flume is typically installed and configured by a system
administrator
Check the Flume documentation if you intend to install it yourself

05-37

Integrating Hadoop Into The Workflow

05-38

FuseDFS and Hoop: Motivation

Many applications generate data which will ultimately reside in
HDFS
If Flume is not an appropriate solution for ingesting the data,
some other method must be used
Typically this is done as a batch process
Problem: many legacy systems do not understand HDFS
Difficult to write to HDFS if the application is not written in Java
May not have Hadoop installed on the system generating the
data
We need some way for these systems to access HDFS

05-39

FuseDFS
FuseDFS is based on FUSE (Filesystem in USEr space)
Allows you to mount HDFS as a regular filesystem
Note: HDFS limitations still exist!
Not intended as a general-purpose filesystem
Files are write-once
Not optimized for low latency
FuseDFS is included as part of the Hadoop distribution

05-40

Hoop
Hoop is an open-source project started at Cloudera
Distributed under the Apache Software License 2.0
Provides an HTTP/HTTPS REST interface to HDFS
Supports both reads and writes from/to HDFS
Can be accessed from within a program
Can be used via command-line tools such as curl or wget
Client accesses the Hoop server
Hoop server then accesses HDFS
Available from http://cloudera.github.com/hoop

REST: REpresentational State Transfer

05-41

Integrating Hadoop Into The Workflow

05-42

Conclusion
In this chapter you have learned
How Hadoop can be integrating into an existing enterprise
How to load data from an existing RDBMS into HDFS using
Sqoop
How to manage real-time data such as log files using Flume
How to access HDFS from legacy systems with FuseDFS and
Hoop

05-43

Chapter 6
Delving Deeper Into
The Hadoop API

06-1

06-2

Delving Deeper Into The Hadoop API

In this chapter you will learn
More about ToolRunner
How to write unit tests with MRUnit
How to specify Combiners to reduce intermediate data
How to use the configure and close methods for Map and
Reduce setup and teardown
How to write custom Partitioners for better load balancing
How to directly access HDFS
How to use the Distributed Cache

06-3

Delving Deeper Into The Hadoop API

More About ToolRunner
Testing With MRUnit
Reducing Intermediate Data with Combiners
The configure and close Methods
Writing Partitioners for Better Load Balancing
Optional Hands-On Exercise
Directly Accessing HDFS
Using The DistributedCache
Hands-On Exercise
Conclusion
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

06-4

Why Use ToolRunner?

In Chapter 4, we used ToolRunner in our driver class
This is not actually required, but is a best practice
ToolRunner uses the GenericOptionsParser class internally
Allows you to specify configuration options on the command line
Also allows you to specify items for the Distributed Cache on the
command line (see later)

06-5

ToolRunner Command Line Options

ToolRunner allows the user to specify configuration options on
the command line
Commonly used to specify Hadoop properties using the D flag
Will override any default or site properties in the configuration
But will not override those set in the driver code
hadoop jar myjar.jar MyDriver D mapred.reduce.tasks=10

Can specify an XML configuration file with -conf

Can specify the default filesystem with -fs uri
Shortcut for D fs.default.name=uri

06-6

Delving Deeper Into The Hadoop API

06-7

An Introduction to Unit Testing

A unit is a small piece of your code
A small piece of functionality
A unit test verifies the correctness of that unit of code
A purist might say that in a well-written unit test, only a single
thing should be able to fail
Generally accepted rule-of-thumb: a unit test should take less
than a second to complete

06-8

Why Write Unit Tests?

Unit testing provides verification that your code is functioning
correctly
Much faster than testing your entire program each time you
modify the code
Fastest MapReduce job on a cluster will take at least 15 seconds
Even in pseudo-distributed mode
Even running in LocalJobRunner mode will take several seconds
LocalJobRunner mode is discussed later in the course
Unit tests help you iterate faster in your code development

06-9

Why MRUnit?
JUnit is a very popular Java unit testing framework
Problem: JUnit cannot be used directly to test Mappers or
Reducers
Unit tests require mocking up InputSplits,
OutputCollector, Reporter,
A lot of work
MRUnit is built on top of JUnit
Provides those mock objects
Allows you to test your code from within an IDE
Much easier to debug

06-10

JUnit Basics
We are using JUnit 4 in class
Earlier versions would also work
@Test
Java annotation
Indicates that this method is a test which JUnit should execute
@Before
Java annotation
Tells JUnit to call this method before every @Test method
Two @Test methods would result in the @Before method
being called twice

06-11

JUnit Basics (contd)

JUnit test methods:
assertEquals(), assertNotNull() etc
Fail if the conditions of the statement are not met
fail(msg)
Fails the test with the given error message
With a JUnit test open in Eclipse, run all tests in the class by
going to Run Run
Eclipse also provides functionality to run all JUnit tests in your
project
Other IDEs have similar functionality

06-12

JUnit: Example Code

import static org.junit.Assert.assertEquals;
import org.junit.Before;
import org.junit.Test;
public class JUnitHelloWorld {
protected String s;
@Before
public void setup() {
s = "HELLO WORLD";
}
@Test
public void testHelloWorldSuccess() {
s = s.toLowerCase();
assertEquals("hello world", s);
}
// will fail even if testHelloWorldSuccess is called first
@Test
public void testHelloWorldFail() {
assertEquals("hello world", s);
}
}

06-13

Using MRUnit to Test MapReduce Code

MRUnit builds on top of JUnit
Provides a mock InputSplit, mock Reporter, mock
OutputCollector and other such classes
Can test just the Mapper, just the Reducer, or the full MapReduce
flow

06-14

MRUnit: Example Code Mapper Unit Test

import hadoop.dev.wordcount.mapred.WordMapper;
import
import
import
import
import
import

org.apache.hadoop.io.IntWritable;
org.apache.hadoop.io.LongWritable;
org.apache.hadoop.io.Text;
org.apache.hadoop.mrunit.MapDriver;
org.junit.Before;
org.junit.Test;

public class MRUnitHelloWorld {

MapDriver<LongWritable, Text, Text, IntWritable> mapDriver;
@Before
public void setUp() {
WordMapper mapper = new WordMapper();
mapDriver = new MapDriver<LongWritable, Text, Text, IntWritable>();
mapDriver.setMapper(mapper);
}
@Test
public void testMapper() {
mapDriver.withInput(new LongWritable(1), new Text("cat dog"));
mapDriver.withOutput(new Text("cat"), new IntWritable(1));
mapDriver.withOutput(new Text("dog"), new IntWritable(1));
mapDriver.runTest();
}
}

06-15

MRUnit: Example Code Mapper Unit

Test(contd)
import hadoop.dev.wordcount.mapred.WordMapper;
import
import
import
import
import
import

org.apache.hadoop.io.IntWritable;
org.apache.hadoop.io.LongWritable;
org.apache.hadoop.io.Text;
org.apache.hadoop.mrunit.MapDriver;
org.junit.Before;
org.junit.Test;

public class MRUnitHelloWorld {

MapDriver<LongWritable, Text, Text, IntWritable> mapDriver;

Import the relevant JUnit classes and the MRUnit

MapDriver class as we will be writing a unit test for
our Mapper

@Before
public void setUp() {
WordMapper mapper = new WordMapper();
mapDriver = new MapDriver<LongWritable, Text, Text, IntWritable>();
mapDriver.setMapper(mapper);
}
@Test
public void testMapper() {
mapDriver.withInput(new LongWritable(1), new Text("cat dog"));
mapDriver.withOutput(new Text("cat"), new IntWritable(1));
mapDriver.withOutput(new Text("dog"), new IntWritable(1));
mapDriver.runTest();
}
}

06-16

MRUnit: Example Code Mapper Unit

Test(contd)
import hadoop.dev.wordcount.mapred.WordMapper;
import
import
import
import
import
import

org.apache.hadoop.io.IntWritable;
org.apache.hadoop.io.LongWritable;
org.apache.hadoop.io.Text;
org.apache.hadoop.mrunit.MapDriver;
org.junit.Before;
org.junit.Test;

public class MRUnitHelloWorld {

MapDriver is an MRUnit class (not a user-defined

driver)

@Test
public void testMapper() {
mapDriver.withInput(new LongWritable(1), new Text("cat dog"));
mapDriver.withOutput(new Text("cat"), new IntWritable(1));
mapDriver.withOutput(new Text("dog"), new IntWritable(1));
mapDriver.runTest();
}
}

06-17

MRUnit: Example Code Mapper Unit

Test(contd)
import hadoop.dev.wordcount.mapred.WordMapper;
import
import
import
import
import
import

org.apache.hadoop.io.IntWritable;
org.apache.hadoop.io.LongWritable;
org.apache.hadoop.io.Text;
org.apache.hadoop.mrunit.MapDriver;
org.junit.Before;
org.junit.Test;

Set up the test. This method will be called before every

test,
just as with JUnit.
class
MRUnitHelloWorld
{

public
MapDriver<LongWritable, Text, Text, IntWritable> mapDriver;

06-18

MRUnit: Example Code Mapper Unit

Test(contd)
import hadoop.dev.wordcount.mapred.WordMapper;
import
import
import
import
import
import

org.apache.hadoop.io.IntWritable;
org.apache.hadoop.io.LongWritable;
org.apache.hadoop.io.Text;
org.apache.hadoop.mrunit.MapDriver;
org.junit.Before;
org.junit.Test;

public class MRUnitHelloWorld {

The test itself.

06-19

MRUnit Drivers
MRUnit has a MapDriver, a ReduceDriver, and a
MapReduceDriver
Methods:
withInput
Specifies input to the Mapper/Reducer
withOutput
Specifies expected output from the Mapper/Reducer
with{Input,Output} support the builder method and can be
chained
add{Input,Output}
Similar to with{Input,Output} but returns void
runTest
Runs the test
run
Runs the test and allows you to retrieve the results
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

06-20

MRUnit Drivers (contd)

Drivers take a single (key, value) pair as input
Can take multiple (key, value) pairs as expected output
If you are calling driver.runTest() multiple times, call
driver.resetOutput() between each call
MRUnit will fail if you do not do this

06-21

Creating Failures, Not Errors

MRUnit throws a RuntimeException when a test fails
In JUnit terms, this means the test has ended as erred rather
than failed
Best practice is to have a failed test be in failed rather than erred
state
This is not required

@Test
public void testMapper() {
mapDriver.withInput(new LongWritable(1), new Text("cat dog"));
mapDriver.withOutput(new Text("cat"), new IntWritable(-1));
mapDriver.withOutput(new Text("dog"), new IntWritable(1));
try {
mapDriver.runTest();
} catch (Exception e) {
fail(e.getMessage());
}
}

06-22

MRUnit Conclusions
You should write unit tests for your code!
As you are performing the Hands-On Exercises in the rest of the
course we strongly recommend that you write unit tests as you
proceed
This will help greatly in debugging your code
Your instructor may ask to see your unit tests if you ask for
assistance!

06-23

Delving Deeper Into The Hadoop API

06-24

The Combiner
Often, Mappers produce large amounts of intermediate data
That data must be passed to the Reducers
This can result in a lot of network traffic
It is often possible to specify a Combiner
Like a mini-Reducer
Runs locally on a single Mappers output
Output from the Combiner is sent to the Reducers
Combiner and Reducer code are often identical
Technically, this is possible if the operation performed is
commutative and associative
In this case, input and output data types for the Combiner/
Reducer must be identical

06-25

MapReduce Example: Word Count

To see how a Combiner works, lets revisit the WordCount
example we covered earlier
map(String input_key, String input_value)
foreach word w in input_value:
emit(w, 1)

reduce(String output_key,
Iterator<int> intermediate_vals)
set count = 0
foreach v in intermediate_vals:
count += v
emit(output_key, count)

06-26

MapReduce Example: Word Count (contd)

Input to the Mapper:
(3414, 'the cat sat on the mat')
(3437, 'the aardvark sat on the sofa')

Output from the Mapper:

('the', 1), ('cat', 1), ('sat', 1), ('on', 1),
('the', 1), ('mat', 1), ('the', 1), ('aardvark', 1),
('sat', 1), ('on', 1), ('the', 1), ('sofa', 1)

06-27

MapReduce Example: Word Count (contd)

Intermediate data sent to the Reducer:
('aardvark', [1])
('cat', [1])
('mat', [1])
('on', [1, 1])
('sat', [1, 1])
('sofa', [1])
('the', [1, 1, 1, 1])

Final Reducer output:

('aardvark', 1)
('cat', 1)
('mat', 1)
('on', 2)
('sat', 2)
('sofa', 1)
('the', 4)

06-28

Word Count With Combiner

A Combiner would reduce the amount of data sent to the
Reducer
Intermediate data sent to the Reducer after a Combiner using the
same code as the Reducer:
('aardvark', [1])
('cat', [1])
('mat', [1])
('on', [2])
('sat', [2])
('sofa', [1])
('the', [4])

Combiners decrease the amount of network traffic required

during the shuffle and sort phase
Often also decrease the amount of work needed to be done by
the Reducer
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

06-29

Specifying a Combiner
To specify the Combiner class to be used in your MapReduce
code, put the following line in your Driver:
conf.setCombinerClass(YourCombinerClass.class);

The Combiner uses the same interface as the Reducer

Takes in a key and a list of values
Outputs zero or more (key, value) pairs
The actual method called is the reduce method in the class
VERY IMPORTANT: Never put code in the Combiner that must be
run as part of your MapReduce job
The Combiner many not be run on the output from some or all of
the Mappers
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

06-30

Delving Deeper Into The Hadoop API

06-31

The configure Method

It is common to want your Mapper or Reducer to execute some
code before the map or reduce method is called
Initialize data structures
Read data from an external file
Set parameters
The configure method is run before the map or reduce method
is called for the first time

public void configure(JobConf conf)

06-32

The close Method

Similarly, you may wish to perform some action(s) after all the
records have been processed by your Mapper or Reducer
The close method is called before the Mapper or Reducer
terminates
public void close() throws IOException

Note that the close() method does not receive the JobConf
object
You could save a reference to the JobConf object in the
configure method and use it in the close method if necessary

06-33

Passing Parameters: The Wrong Way!

public class MyClass {
private static int param;
...
private static class MyMapper extends MapReduceBase ... {
public void map... {
int v = param;
}
}
...
public static void main(String[] args) throws IOException {
JobConf conf = new JobConf(MyClass.class);
param = 5;
...
JobClient.runJob(conf);
}
}

06-34

Passing Parameters: The Right Way

public class MyClass {
...
private static class MyMapper extends MapReduceBase ... {
public void configure(JobConf job) {
int v = job.getInt("param", 0);
...
}
...
public void map...
}
...
public static void main(String[] args) throws IOException {
JobConf conf = new JobConf(MyClass.class);
conf.setInt("param", 5)
...
JobClient.runJob(conf);
}
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

06-35

Delving Deeper Into The Hadoop API

06-36

What Does The Partitioner Do?

The Partitioner divides up the keyspace
Controls which Reducer each intermediate key and its associated
values goes to
Often, the default behavior is fine
Default is the HashPartitioner
public class HashPartitioner<K2, V2> implements Partitioner<K2, V2> {
public void configure(JobConf job) {}
public int getPartition(K2 key, V2 value,
int numReduceTasks) {
return (key.hashCode() & Integer.MAX_VALUE) % numReduceTasks;
}
}

06-37

Custom Partitioners
Sometimes you will need to write your own Partitioner
Example: your key is a custom WritableComparable which
contains a pair of values (a, b)
You may decide that all keys with the same value for a need to go
to the same Reducer
The default Partitioner is not sufficient in this case

06-38

Custom Partitioners (contd)

Custom Partitioners are needed when performing a secondary
sort (see later)
Custom Partitioners are also useful to avoid potential
performance issues
To avoid one Reducer having to deal with many very large lists of
values
Example: in our word count job, we wouldn't want a single reduce
dealing with all the three- and four-letter words, while another
only had to handle 10- and 11-letter words

06-39

Creating a Custom Partitioner

To create a custom partitioner:
1. Create a class for the partitioner
Should implement the Partitioner interface
2. Create a method in the class called getPartition
Receives the key, the value, and the number of Reducers
Should return an int between 0 and one less than the number of
Reducers
e.g., if it is told there are 10 Reducers, it should return an int
between 0 and 9
3. Specify the custom partitioner in your driver code
conf.setPartitionerClass(MyPartitioner.class);

06-40

Delving Deeper Into The Hadoop API

06-41

Optional Hands-On Exercise

In this optional Hands-On Exercise, you will write code which
uses a Partitioner and multiple Reducers
Note: Depending on time, your instructor may omit this exercise,
or may choose to do it later in the week
Please refer to the Hands-On Exercise Manual

06-42

Delving Deeper Into The Hadoop API

06-43

Accessing HDFS Programmatically

In addition to using the command-line shell, you can access
HDFS programmatically
Useful if your code needs to read or write side data in addition to
the standard MapReduce inputs and outputs
Beware: HDFS is not a general-purpose filesystem!
Files cannot be modified once they have been written, for
example
Hadoop provides the FileSystem abstract base class
Provides an API to generic file systems
Could be HDFS
Could be your local file system
Could even be, for example, Amazon S3

06-44

The FileSystem API

In order to use the FileSystem API, retrieve an instance of it
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(conf);

The conf object has read in the Hadoop configuration files, and
therefore knows the address of the NameNode etc.
A file in HDFS is represented by a Path object
Path p = new Path("/path/to/my/file");

06-45

The FileSystem API (contd)

Some useful API methods:
FSDataOuputStream create(...)
Extends java.io.DataOutputStream
Provides methods for writing primitives, raw bytes etc
FSDataInputStream open(...)
Extends java.io.DataInputStream
Provides methods for reading primitives, raw bytes etc
boolean delete(...)
boolean mkdirs(...)
void copyFromLocalFile(...)
void copyToLocalFile(...)
FileStatus[] listStatus(...)

06-46

The FileSystem API: Directory Listing

Get a directory listing:
Path p = new Path("/my/path");
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(conf);
FileStatus[] fileStats = fs.listStatus(p);
for (int i = 0; i < fileStats.length; i++) {
Path f = fileStats[i].getPath();
// do something interesting
}

06-47

The FileSystem API: Writing Data

Write data to a file
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(conf);
Path p = new Path("/my/path/foo");
FSDataOutputStream out = fs.create(path, false);
// write some raw bytes
out.write(getBytes());
// write an int
out.writeInt(getInt());
...
out.close();
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

06-48

Delving Deeper Into The Hadoop API

06-49

The Distributed Cache: Motivation

A common requirement is for a Mapper or Reducer to need
access to some side data
Lookup tables
Dictionaries
Standard configuration values
One option: read directly from HDFS in the configure method
Works, but is not scalable
The DistributedCache provides an API to push data to all slave
nodes
Transfer happens behind the scenes before any task is executed
Note: DistributedCache is read-only
Files in the DistributedCache are automatically deleted from slave
nodes when the job finishes
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

06-50

Using the DistributedCache: The Difficult Way

Place the files into HDFS
Configure the DistributedCache in your driver code
JobConf job = new JobConf();
DistributedCache.addCacheFile(new URI("/myapp/lookup.dat"), job);
DistributedCache.addFileToClassPath(new Path("/myapp/mylib.jar"), job);
DistributedCache.addCacheArchive(new URI("/myapp/map.zip", job);
DistributedCache.addCacheArchive(new URI("/myapp/mytar.tar", job);
DistributedCache.addCacheArchive(new URI("/myapp/mytgz.tgz", job);
DistributedCache.addCacheArchive(new URI("/myapp/mytargz.tar.gz", job);

.jar files added with addFileToClassPath will be added to

your Mapper or Reducers classpath
Files added with addCacheArchive will automatically be
dearchived/decompressed
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

06-51

Using the DistributedCache: The Easy Way

If you are using ToolRunner, you can add files to the
DistributedPath directly from the command line when you run
the job
No need to copy the files to HDFS first
Use the -files option to add files
hadoop jar myjar.jar MyDriver -files file1, file2, file3, ...

The -archives flag adds archived files, and automatically

unarchives them on the destination machines
The -libjars flag adds jar files to the classpath

06-52

Accessing Files in the DistributedCache

Files added to the DistributedCache are made available in your
tasks local working directory
Access them from your Mapper or Reducer the way you would
read any ordinary local file
File f = new File("file_name_here");

06-53

Delving Deeper Into The Hadoop API

06-54

Hands-On Exercise
In this Hands-On Exercise, you will gain practice writing
combiners and creating unit tests
Please refer to the Hands-On Exercise Manual

06-55

Delving Deeper Into The Hadoop API

06-56

Conclusion
In this chapter you have learned
More about ToolRunner
How to write unit tests with MRUnit
How to specify Combiners to reduce intermediate data
How to use the configure and close methods for Map and
Reduce setup and teardown
How to write custom Partitioners for better load balancing
How to directly access HDFS
How to use the Distributed Cache

06-57

Chapter 7
Common MapReduce
Algorithms

07-1

07-2

Common MapReduce Algorithms

In this chapter you will learn
Some typical MapReduce algorithms, including
Sorting
Searching
Indexing
Term Frequency Inverse Document Frequency
Word Co-Occurrence
We will also briefly discuss machine learning with Hadoop

07-3

Common MapReduce Algorithms

Introduction
Sorting and Searching
Indexing
Machine Learning
TF-IDF
Word Co-Occurrence
Hands-On Exercise
Conclusion
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

07-4

Introduction
MapReduce jobs tend to be relatively short in terms of lines of
code
It is typical to combine multiple small MapReduce jobs together
in a single workflow
Often using Oozie (see later)
You are likely to find that many of your MapReduce jobs use very
similar code
In this chapter we present some very common MapReduce
algorithms
These algorithms are frequently the basis for more complex
MapReduce jobs

07-5

Common MapReduce Algorithms

07-6

Sorting
MapReduce is very well suited to sorting large data sets
Recall: keys are passed to the reducer in sorted order
Assuming the file to be sorted contains lines with a single value:
Mapper is merely the identity function for the value
(k, v) -> (v, _)
Reducer is the identity function
(k, _) -> (k, '')

07-7

Sorting (contd)
Trivial with a single reducer
For multiple reducers, need to choose a partitioning function
such that if k1 < k2, partition(k1) <= partition(k2)

07-8

Sorting as a Speed Test of Hadoop

Sorting is frequently used as a speed test for a Hadoop cluster
Mapper and Reducer are trivial
Therefore sorting is effectively testing the Hadoop
frameworks I/O
Good way to measure the increase in performance if you enlarge
your cluster
Run and time a sort job before and after you add more nodes
terasort is one of the sample jobs provided with Hadoop
Creates and sorts very large files

07-9

Searching
Assume the input is a set of files containing lines of text
Assume the Mapper has been passed the pattern for which to
search as a special parameter
We saw how to pass parameters to your Mapper in the previous
chapter
Algorithm:
Mapper compares the line against the pattern
If the pattern matches, Mapper outputs (line, _)
Or (filename+line, _), or
If the pattern does not match, Mapper outputs nothing
Reducer is the Identity Reducer
Just outputs each intermediate key

07-10

Common MapReduce Algorithms

07-11

Indexing
Assume the input is a set of files containing lines of text
Key is the byte offset of the line, value is the line itself
We can retrieve the name of the file using the Reporter object
More details on how to do this later

07-12

Inverted Index Algorithm

Mapper:
For each word in the line, emit (word, filename)
Reducer:
Identity function
Collect together all values for a given key (i.e., all filenames
for a particular word)
Emit (word, filename_list)

07-13

Inverted Index: Dataflow

07-14

Aside: Word Count

Recall the WordCount example we used earlier in the course
For each word, Mapper emitted (word, 1)
Very similar to the inverted index
This is a common theme: reuse of existing Mappers, with minor
modifications

07-15

Common MapReduce Algorithms

07-16

Machine Learning: Introduction

Machine Learning is a complex discipline
Much research is ongoing
Here we merely give a very high-level overview of some aspects
of ML

07-17

What Is Machine Learning Not?

Most programs tell computers exactly what to do
Database transactions and queries
Controllers
Phone systems, manufacturing processes, transport,
weaponry, etc.
Media delivery
Simple search
Social systems
Chat, blogs, e-mail etc.

07-18

What Is Machine Learning?

An alternative technique is to have computers learn what to do
Machine Learning refers to a few classes of program that
leverage collected data to drive future program behavior
This represents another major opportunity to gain value from
data

07-19

Why Use Hadoop for Machine Learning?

Machine Learning systems are sensitive to the skill you bring to
them
However, practitioners often agree [Banko and Brill, 2001]:

Its not who has the best algorithms that wins.

Its who has the most data.
or
Theres no data like more data.
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

07-20

The Three Cs
Machine Learning is an active area of research and new
applications
There are three well-established categories of techniques for
exploiting data:
Collaborative filtering (recommendations)
Clustering
Classification

07-21

Collaborative Filtering
Collaborative Filtering is a technique for recommendations
Example application: given people who each like certain books,
learn to suggest what someone may like based on what they
already like
Very useful in helping users navigate data by expanding to topics
that have affinity with their established interests
Collaborative Filtering algorithms are agnostic to the different
types of data items involved
So they are equally useful in many different domains

07-22

Clustering
Clustering algorithms discover structure in collections of data
Where no formal structure previously existed
They discover what clusters, or groupings, naturally occur in
data
Examples:
Finding related news articles
Computer vision (groups of pixels that cohere into objects)

07-23

Classification
The previous two techniques are considered unsupervised
learning
The algorithm discovers groups or recommendations itself
Classification is a form of supervised learning
A classification system takes a set of data records with known
labels
Learns how to label new records based on that information
Example:
Given a set of e-mails identified as spam/not spam, label new emails as spam/not spam
Given tumors identified as benign or malignant, classify new
tumors

07-24

Mahout: A Machine Learning Library

Mahout is a Machine Learning library
Included in CDH3
Contains algorithms for each of the categories listed
Algorithms included in Mahout:
Recommendation

Clustering

Classification

Pearson correlation
Log likelihood
Spearman correlation
Tanimoto coefficient
Singular value
decomposition (SVD)
Llinear interpolation
Cluster-based
recommenders

k-means clustering
Canopy clustering
Fuzzy k-means
Latent Dirichlet
analysis (LDA)

Stochastic gradient
descent (SGD)
Support vector
machine (SVM)
Nave Bayes
Complementary nave
Bayes
Random forests

07-25

Common MapReduce Algorithms

07-26

Term Frequency Inverse Document

Frequency
Term Frequency Inverse Document Frequency (TF-IDF)
Answers the question How important is this term in a document
Known as a term weighting function
Assigns a score (weight) to each term (word) in a document
Very commonly used in text processing and search
Has many applications in data mining

07-27

TF-IDF: Motivation
Merely counting the number of occurrences of a word in a
document is not a good enough measure of its relevance
If the word appears in many other documents, it is probably less
relevance
Some words appear too frequently in all documents to be
relevant
Known as stopwords
TF-IDF considers both the frequency of a word in a given
document and the number of documents which contain the word

07-28

TF-IDF: Data Mining Example

Consider a music recommendation system
Given many users music libraries, provide you may also like
suggestions
If user A and user B have similar libraries, user A may like an
artist in user Bs library
But some artists will appear in almost everyones library, and
should therefore be ignored when making recommendations
Almost everyone has The Beatles in their record collection!

07-29

TF-IDF Formally Defined

Term Frequency (TF)
Number of times a term appears in a document (i.e., the count)
Inverse Document Frequency (IDF)

"N%
idf = log$ '
#n&
N: total number of documents
n: number of documents that contain a term

TF-IDF
TF IDF

07-30

Computing TF-IDF
What we need:
Number of times t appears in a document
Different value for each document
Number of documents that contains t
One value for each term
Total number of documents
One value

07-31

Computing TF-IDF With MapReduce

Overview of algorithm: 3 MapReduce jobs
Job 1: compute term frequencies
Job 2: compute number of documents each word occurs in
Job 3: compute TF-IDF
Notation in following slides:
tf = term frequency
n = number of documents a term appears in
N = total number of documents
docid = a unique id for each document

07-32

Computing TF-IDF: Job 1 Compute tf

Mapper
Input: (docid, contents)
For each term in the document, generate a (term, docid) pair
i.e., we have seen this term in this document once
Output: ((term, docid), 1)
Reducer
Sums counts for word in document
Outputs ((term, docid), tf)
I.e., the term frequency of term in docid is tf
We can add a Combiner, which will use the same code as the
Reducer

07-33

Computing TF-IDF: Job 2 Compute n

Mapper
Input: ((term, docid), tf)
Output: (term, (docid, tf, 1))
Reducer
Sums 1s to compute n (number of documents containing term)
Note: need to buffer (docid, tf) pairs while we are doing this
(more later)
Outputs ((term, docid), (tf, n))

07-34

Computing TF-IDF: Job 3 Compute TF-IDF

Mapper
Input: ((term, docid), (tf, n))
Assume N is known (easy to find)
Output ((term, docid), TF IDF)
Reducer
The identity function

07-35

Computing TF-IDF: Working At Scale

Job 2: We need to buffer (docid, tf) pairs counts while summing
1s (to compute n)
Possible problem: pairs may not fit in memory!
How many documents does the word the occur in?
Possible solutions
Ignore very-high-frequency words
Write out intermediate data to a file
Use another MapReduce pass

07-36

TF-IDF: Final Thoughts

Several small jobs add up to full algorithm
Thinking in MapReduce often means decomposing a complex
algorithm into a sequence of smaller jobs
Beware of memory usage for large amounts of data!
Any time when you need to buffer data, theres a potential
scalability bottleneck

07-37

Common MapReduce Algorithms

07-38

Word Co-Occurrence: Motivation

Word Co-Occurrence measures the frequency with which two
words appear close to each other in a corpus of documents
For some definition of close
This is at the heart of many data-mining techniques
Provides results for people who did this, also do that
Examples:
Shopping recommendations
Credit risk analysis
Identifying people of interest

07-39

Word Co-Occurrence: Algorithm

Mapper
map(docid a, doc d) {
foreach w in d do
foreach u near w do
emit(pair(w, u), 1)
}

Reducer
reduce(pair p, Iterator counts) {
s = 0
foreach c in counts do
s += c
emit(p, s)
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

07-40

Common MapReduce Algorithms

07-41

Hands-On Exercise: Inverted Index

In this Hands-On Exercise, you will write a MapReduce program
to generate an inverted index of a set of documents
Please refer to the Hands-On Exercise Instructions

07-42

Common MapReduce Algorithms

07-43

Conclusion
In this chapter you have learned
Some typical MapReduce algorithms, including
Sorting
Searching
Indexing
Term Frequency Inverse Document Frequency
Word Co-Occurrence
We also briefly discussed machine learning with Hadoop

07-44

Chapter 8
An Introduction to
Hive and Pig

08-1

08-2

An Introduction to Hive and Pig

In this chapter you will learn
What features Hive provides
What features Pig provides

08-3

An Introduction to Hive and Pig

Motivation for Hive and Pig
Hive Basics
Pig Basics
Which To Use?
Hands-On Exercise
Conclusion

08-4

Hive and Pig: Motivation

MapReduce code is typically written in Java
Although it can be written in other languages using Hadoop
Streaming
Requires:
A programmer
Who is a good Java programmer
Who understands how to think in terms of MapReduce
Who understands the problem theyre trying to solve
Who has enough time to write and test the code
Who will be available to maintain and update the code in the
future as requirements change

08-5

Hive and Pig: Motivation (contd)

Many organizations have only a few developers who can write
good MapReduce code
Meanwhile, many other people want to analyze data
Business analysts
Data scientists
Statisticians
Data analysts
Whats needed is a higher-level abstraction on top of MapReduce
Providing the ability to query the data without needing to know
MapReduce intimately
Hive and Pig address these needs

08-6

An Introduction to Hive and Pig

Motivation for Hive and Pig
Hive Basics
Pig Basics
Which To Use?
Hands-On Exercise
Conclusion

08-7

Hive: Introduction
Hive was originally developed at Facebook
Provides a very SQL-like language
Can be used by people who know SQL
Under the covers, generates MapReduce jobs that run on the
Hadoop cluster
Enabling Hive requires almost no extra work by the system
administrator

08-8

The Hive Data Model

Hive layers table definitions on top of data in HDFS
Tables
Typed columns (int, float, string, boolean and so on)
Also, list: map (for JSON-like data)
Partitions
e.g., to range-partition tables by date
Buckets
Hash partitions within ranges (useful for sampling, join
optimization)

08-9

Hive Datatypes
Primitive types:
TINYINT
INT
BIGINT
BOOLEAN
DOUBLE
STRING
Type constructors:
ARRAY < primitive-type >
MAP < primitive-type, data-type >
STRUCT < col-name : data-type, ... >

08-10

The Hive Metastore

Hives Metastore is a database containing table definitions and
other metadata
By default, stored locally on the client machine in a Derby
database
If multiple people will be using Hive, the system administrator
should create a shared Metastore
Usually in MySQL or some other relational database server

08-11

Hive Data: Physical Layout

Hive tables are stored in Hives warehouse directory in HDFS
By default, /user/hive/warehouse
Tables are stored in subdirectories of the warehouse directory
Partitions form subdirectories of tables
Possible to create external tables if the data is already in HDFS
and should not be moved from its current location
Actual data is stored in flat files
Control character-delimited text, or SequenceFiles
Can be in arbitrary format with the use of a custom Serializer/
Deserializer (SerDe)

08-12

Starting The Hive Shell

To launch the Hive shell, start a terminal and run
$ hive

Results in the Hive prompt:

hive>

08-13

Hive Basics: Creating Tables

hive> SHOW TABLES;

hive> CREATE TABLE shakespeare
(freq INT, word STRING)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY '\t'
STORED AS TEXTFILE;
hive> DESCRIBE shakespeare;

08-14

Loading Data Into Hive

Data is loaded into Hive with the LOAD DATA INPATH statement
Assumes that the data is already in HDFS
LOAD DATA INPATH "shakespeare_freq" INTO TABLE shakespeare;

If the data is on the local filesystem, use LOAD DATA LOCAL

INPATH
Automatically loads it into HDFS

08-15

Basic SELECT Queries

Hive supports most familiar SELECT syntax
hive> SELECT * FROM shakespeare LIMIT 10;
hive> SELECT * FROM shakespeare
WHERE freq > 100 ORDER BY freq ASC
LIMIT 10;

08-16

Joining Tables
Joining datasets is a complex operation in standard Java
MapReduce
We will cover this later in the course
In Hive, its easy!
SELECT s.word, s.freq, k.freq FROM
shakespeare s JOIN kjv k ON
(s.word = k.word)
WHERE s.freq >= 5;

08-17

Storing Output Results

The SELECT statement on the previous slide would write the data
to the console
To store the results in HDFS, create a new table then write, for
example:
INSERT OVERWRITE TABLE newTable
SELECT s.word, s.freq, k.freq FROM
shakespeare s JOIN kjv k ON
(s.word = k.word)
WHERE s.freq >= 5;

Results are stored in the table

Results are just files within the newTable directory
Data can be used in subsequent queries, or in MapReduce
jobs
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

08-18

Creating User-Defined Functions

Hive supports manipulation of data via user-created functions
Example:
INSERT OVERWRITE TABLE u_data_new
SELECT
TRANSFORM (userid, movieid, rating, unixtime)
USING 'python weekday_mapper.py'
AS (userid, movieid, rating, weekday)
FROM u_data;

08-19

Hive Limitations
Not all standard SQL is supported
No correlated subqueries, for example
No support for UPDATE or DELETE
No support for INSERTing single rows
Relatively limited number of built-in functions
No datatypes for date or time
Use the STRING datatype instead

08-20

Hive: Where To Learn More

Main Web site is at http://hive.apache.org/
Cloudera training course: Analyzing Data With Hive And Pig

08-21

An Introduction to Hive and Pig

Motivation for Hive and Pig
Hive Basics
Pig Basics
Which To Use?
Hands-On Exercise
Conclusion

08-22

Pig: Introduction
Pig was originally created at Yahoo! to answer a similar need to
Hive
Many developers did not have the Java and/or MapReduce
knowledge required to write standard MapReduce programs
But still needed to query data
Pig is a dataflow language
Language is called PigLatin
Relatively simple syntax
Under the covers, PigLatin scripts are turned into MapReduce
jobs and executed on the cluster

08-23

Pig Installation
Installation of Pig requires no modification to the cluster
The Pig interpreter runs on the client machine
Turns PigLatin into standard Java MapReduce jobs, which are
then submitted to the JobTracker
There is (currently) no shared metadata, so no need for a shared
metastore of any kind

08-24

Pig Concepts
In Pig, a single element of data is an atom
A collection of atoms such as a row, or a partial row is a tuple
Tuples are collected together into bags
Typically, a PigLatin script starts by loading one or more datasets
into bags, and then creates new bags by modifying those it
already has

08-25

Pig Features
Pig supports many features which allow developers to perform
sophisticated data analysis without having to write Java
MapReduce code
Joining datasets
Grouping data
Referring to elements by position rather than name
Useful for datasets with many elements
Loading non-delimited data using a custom SerDe
Creation of user-defined functions, written in Java
And more

08-26

A Sample Pig Script

emps = LOAD 'people.txt' AS (id, name, salary);
rich = FILTER emps BY salary > 100000;
srtd = ORDER rich BY salary DESC;
STORE srtd INTO 'rich_people';

Here, we load a file into a bag called emps

Then we create a new bag called rich which contains just those
records where the salary portion is greater than 100000
Finally, we write the contents of the srtd bag to a new directory
in HDFS
By default, the data will be written in tab-separated format
Alternatively, to write the contents of a bag to the screen, say
DUMP srtd;
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

08-27

More PigLatin
To view the structure of a bag:
DESCRIBE bagname;

Joining two datasets:

data1
data2
jnd =
STORE

= LOAD 'data1' AS (col1, col2, col3, col4);

= LOAD 'data2' AS (colA, colB, colC);
JOIN data1 BY col3, data2 BY colA;
jnd INTO 'outfile';

08-28

More PigLatin: Grouping

Grouping:
grpd = GROUP bag1 BY elementX

Creates a new bag

Each tuple in grpd has an element called group, and an
element called bag1
The group element has a unique value for elementX from bag1
The bag1 element is itself a bag, containing all the tuples from
bag1 with that value for elementX

08-29

More PigLatin: FOREACH

The FOREACH...GENERATE statement iterates over members of a
bag
Example:
justnames = FOREACH emps GENERATE name;

Can combine with COUNT:

summedUp = FOREACH grpd GENERATE group,
COUNT(bag1) AS elementCount;

08-30

Pig: Where To Learn More

Main Web site is at http://pig.apache.org/
Follow the links on the left-hand side of the page to
Documentation, then Release 0.7.0, then Pig Latin 1 and Pig
Latin 2
Cloudera training course: Analyzing Data With Hive And Pig

08-31

An Introduction to Hive and Pig

Motivation for Hive and Pig
Hive Basics
Pig Basics
Which To Use?
Hands-On Exercise
Conclusion

08-32

Choosing Between Pig and Hive

Typically, organizations wanting an abstraction on top of
standard MapReduce will choose to use either Hive or Pig
Which one is chosen depends on the skillset of the target users
Those with an SQL background will naturally gravitate towards
Hive
Those who do not know SQL will often choose Pig
Each has strengths and weaknesses; it is worth spending some
time investigating each so you can make an informed decision
Some organizations are now choosing to use both
Pig deals better with less-structured data, so Pig is used to
manipulate the data into a more structured form, then Hive is
used to query that structured data

08-33

An Introduction to Hive and Pig

Motivation for Hive and Pig
Hive Basics
Pig Basics
Which To Use?
Hands-On Exercise
Conclusion

08-34

Hands-On Exercise: Manipulating Data with

Hive or Pig
In this Hands-On Exercise, you will manipulate a dataset using
Hive
Please refer to the Hands-On Exercise Manual

08-35

An Introduction to Hive and Pig

Motivation for Hive and Pig
Hive Basics
Pig Basics
Which To Use?
Hands-On Exercise
Conclusion

08-36

Conclusion
In this chapter you have learned
What features Hive provides
What features Pig provides

08-37

Chapter 9
Practical Development
Tips and Techniques

09-1

09-2

Practical Development Tips and Techniques

In this chapter you will learn
Strategies for debugging your MapReduce code
How to write and subsequently view log files
How to retrieve job information with Counters
What issues to consider when using file compression
How to determine the optimal number of Reducers for a job
How to create Map-only MapReduce jobs

09-3

Practical Development Tips and Techniques

Strategies for Debugging MapReduce Code
Using LocalJobRunner for Easier Debugging
Logging
Retrieving Job Information with Counters
Splittable File Formats
Determining the Optimal Number of Reducers
Map-Only MapReduce Jobs
Hands-On Exercise
Conclusion
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

09-4

Introduction to Debugging
Debugging MapReduce code is difficult!
Each instance of a Mapper runs as a separate task
Often on a different machine
Difficult to attach a debugger to the process
Difficult to catch edge cases
Very large volumes of data mean that unexpected input is likely
to appear
Code which expects all data to be well-formed is likely to fail

09-5

Common-Sense Debugging Tips

Code defensively
Ensure that input data is in the expected format
Expect things to go wrong
Catch exceptions
Start small, build incrementally
Make as much of your code as possible Hadoop-agnostic
Makes it easier to test
Write unit tests
Test locally whenever possible
With small amounts of data
Then test in pseudo-distributed mode
Finally, test on the cluster
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

09-6

Basic Debugging Strategies

You can throw exceptions if a particular condition is met
For example, if illegal data is found
throw new RuntimeException("Your message here");

This causes the task to fail

If a task fails four times, the entire job will fail

09-7

Basic Debugging Strategies (contd)

If you suspect the input data of being faulty, you may be tempted
to log the (key, value) pairs your Mapper receives
Reasonable for small amounts of input data
Caution! If your job runs across 500GB of input data, you could
be writing 500GB of log files!
Remember to think at scale
Instead, wrap vulnerable sections of code in try {...} blocks
Write logs in the catch {...} block
This way only critical data is logged

09-8

Testing Strategies
When testing in pseudo-distributed mode, ensure that you are
testing with a similar environment to that on the real cluster
Same amount of RAM allocated to the task JVMs
Same version of Hadoop
Same version of Java
Same versions of third-party libraries

09-9

Practical Development Tips and Techniques

09-10

Testing Locally
Hadoop can run MapReduce in a single, local process
Does not require any Hadoop daemons to be running
Uses the local filesystem
Known as the LocalJobRunner
This is a very useful way of quickly testing incremental changes
to code
To run in LocalJobRunner mode, add the following lines to your
driver code:
conf.set("mapred.job.tracker", "local");
conf.set("fs.default.name", "file:///");

Or set these options on the command line with the -D flag

09-11

Testing Locally (contd)

Some limitations of LocalJobRunner mode:
DistributedCache does not work
The job can only specify a single Reducer
Some beginner mistakes may not be caught
For example, attempting to share data between Mappers will
work, because the code is running in a single JVM

09-12

LocalJobRunner Mode in Eclipse

Eclipse can be configured to run Hadoop code in
LocalJobRunner mode
From within the IDE
This allows very rapid development iterations
Agile programming

09-13

Practical Development Tips and Techniques

09-14

Before Logging: stdout and stderr

Tried-and-true debugging technique: write to stdout or stderr
If running in LocalJobRunner mode, you will see the results of
System.err.println()
If running on a cluster, that output will not appear on your
console
Output is visible via Hadoops Web UI

09-15

Aside: The Hadoop Web UI

All Hadoop daemons contain a Web server
Exposes information on a well-known port
Most important for developers is the JobTracker Web UI
http://<job_tracker_address>:50030/
http://localhost:50030/ if running in pseudo-distributed
mode
Also useful: the NameNode Web UI
http://<name_node_address>:50070/

09-16

Aside: The Hadoop Web UI (contd)

Your instructor will now demonstrate the JobTracker UI

09-17

Logging: Better Than Printing

println statements rapidly become awkward
Turning them on and off in your code is tedious, and leads to
errors
Logging provides much finer-grained control over:
What gets logged
When something gets logged
How something is logged

09-18

Logging With log4j

Hadoop uses log4j to generate all its log files
Your Mappers and Reducers can also use log4j
All the initialization is handled for you by Hadoop
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
class FooMapper implements Mapper {
public static final Log LOGGER =
LogFactory.getLog(FooMapper.class.getName());
...
}

09-19

Logging With log4j (contd)

Simply send strings to loggers tagged with severity levels:
LOGGER.trace("message");
LOGGER.debug("message");
LOGGER.info("message");
LOGGER.warn("message");
LOGGER.error("message"):

Beware expensive operations like concatenation

To avoid performance penalty, make it conditional like this:
if (LOGGER.isDebugEnabled()) {
LOGGER.debug("Account info:" + acct.getReport());
}

09-20

log4j Configuration
Configuration for log4j is stored in
/etc/hadoop/conf/log4j.properties
Can change global log settings with hadoop.root.log property
Can override log level on a per-class basis:
log4j.logger.org.apache.hadoop.mapred.JobTracker=WARN
log4j.logger.org.apache.hadoop.mapred.FooMapper=DEBUG

Programmatically:
LOGGER.setLevel(Level.WARN);

09-21

Dynamically Setting Log Levels

Although log levels can be set in log4j.properties, this would
require modification of files on all slave nodes
In practice, this is unrealistic
Instead, a good solution is to set the log level in your code based
on a command-line parameter

09-22

Dynamically Setting Log Levels (contd)

In the code for your Mapper or Reducer:
public void configure(JobConf conf) {
if ("DEBUG".equals(conf.get("com.cloudera.job.logging")){
LOGGER.setLevel(Level.DEBUG);
LOGGER.debug("** Log Level set to DEBUG **");
}
}

Then on the command line, specify the log level:

$ hadoop jar wc.jar WordCountWTool \
Dcom.cloudera.job.logging=DEBUG myclass indir outdir

09-23

Where Are Log Files Stored?

Log files are stored by default at
/var/log/hadoop/userlogs/${task.id}/syslog
on the machine where the task attempt ran
Configurable
Tedious to have to ssh in to a node to view its logs
Much easier to use the JobTracker Web UI
Automatically retrieves and displays the log files for you

09-24

Practical Development Tips and Techniques

09-25

What Are Counters?

Counters provide a way for Mappers or Reducers to pass
aggregate values back to the driver after the job has completed
Their values are also visible from the JobTrackers Web UI
And are reported on the console when the job ends
Very basic: just have a name and a value
Value can be incremented within the code
Counters are collected into Groups
Within the group, each Counter has a name
Example: A group of Counters called RecordType
Names: TypeA, TypeB, TypeC
Appropriate Counter will be incremented as each record is read in
the Mapper

09-26

What Are Counters? (contd)

Counters provide a way for Mappers or Reducers to pass
aggregate values back to the driver after the job has completed
Their values are also visible from the JobTrackers Web UI
Counters can be set and incremented via the method
Reporter.incrCounter(group, name, amount);

Example:
r.incrCounter("RecordType", "TypeA", 1);

09-27

Retrieving Counters in the Driver Code

To retrieve Counters in the Driver code after the job is complete,
use code like this:
RunningJob job = JobClient.runJob(conf);
long typeARecords =
job.getCounters().findCounter("RecordType", "TypeA").getValue();
long typeBRecords =
job.getCounters().findCounter("RecordType", "TypeB").getValue();

09-28

Counters: Caution
Do not rely on a counters value from the Web UI while a job is
running
Due to possible speculative execution, a counters value could
appear larger than the actual final value
Modifications to counters from subsequently killed/failed tasks will
be removed from the final count

09-29

Practical Development Tips and Techniques

09-30

InputSplits: Reprise
Recall that each Mapper processes a single InputSplit
InputSplits are calculated by the client program before the job is
submitted to the JobTracker
Typically, an InputSplit equates to an HDFS block
So the number of Mappers will equal the number of HDFS blocks
in the input data

09-31

Hadoop and Compressed Files

Hadoop understands a variety of file compression formats
Including GZip
If a compressed file is included as one of the files to be
processed, Hadoop will automatically decompress it and pass
the decompressed contents to the Mapper
There is no need for the developer to worry about decompressing
the file
However, GZip is not a splittable file format
A GZipped file can only be decompressed by starting at the
beginning of the file and continuing on to the end
You cannot start decompressing the file part of the way through it

09-32

Non-Splittable File Formats and Hadoop

If the MapReduce framework receives a non-splittable file (such
as a GZipped file) it passes the entire file to a single Mapper
This can result in one Mapper running for far longer than the
others
It is dealing with an entire file, while the others are dealing with
smaller portions of files
Speculative execution can occur
Although this will provide no benefit
Typically it is not a good idea to use GZip to compress files
which will be processed by MapReduce

09-33

Splittable Compression Formats: LZO

One splittable compression format is LZO
Because of licensing restrictions, LZO cannot be shipped with
Hadoop
But it is easy to add
See https://github.com/cloudera/hadoop-lzo
LZO files have set points within the file at which compression
can begin
InputSplits are calculated based on those points
The point closest to each block boundary is chosen

09-34

Splittable Compression Formats: Snappy

Snappy is a relatively new compression format
Developed at Google
Splittable
Very fast
Snappy does not compress a file and produce, e.g., a file with
a .snappy extension
Instead, it compresses data within a file
That data can be decompressed automatically by Hadoop (or
other programs) when the file is read
Works well with SequenceFiles, Avro files
Snappy is still new to Hadoop, but is already seeing significant
adoption

09-35

Practical Development Tips and Techniques

09-36

How Many Reducers Do You Need?

An important consideration when creating your job is to
determine the number of Reducers specified
Default is a single Reducer
With a single Reducer, one task receives all keys in sorted order
This is sometimes advantageous if the output must be in
completely sorted order
Can cause significant problems if there is a large amount of
intermediate data
Node on which the Reducer is running may not have enough
disk space to hold all intermediate data
The Reducer will take a long time to run

09-37

Jobs Which Require a Single Reducer

If a job needs to output a file where all keys are listed in sorted
order, a single Reducer must be used
Alternatively, the TotalOrderPartitioner can be used
Uses an externally generated file which contains information
about intermediate key distribution
Partitions data such that all keys which go to the first reducer are
smaller than any which go to the second, etc
In this way, multiple Reducers can be used
Concatenating the Reducers output files results in a totally
ordered list

09-38

Jobs Which Require a Fixed Number of

Reducers
Some jobs will require a specific number of Reducers
Example: a job must output one file per day of the week
Key will be the weekday
Seven Reducers will be specified
A Partitioner will be written which sends one key to each
Reducer

09-39

Jobs With a Variable Number of Reducers

Many jobs can be run with a variable number of Reducers
Developer must decide how many to specify
Each Reducer should get a reasonable amount of intermediate
data, but not too much
Chicken-and-egg problem
Typical way to determine how many Reducers to specify:
Test the job with a relatively small test data set
Extrapolate to calculate the amount of intermediate data
expected from the real input data
Use that to calculate the number of Reducers which should be
specified

09-40

Practical Development Tips and Techniques

09-41

Map-Only MapReduce Jobs

There are many types of job where only a Mapper is needed
Examples:
Image processing
File format conversion
Input data sampling
ETL

09-42

Creating Map-Only Jobs

To create a Map-only job, set the number of Reducers to 0 in your
Driver code
conf.setNumReduceTasks(0);

Anything written using the OutputCollector.collect method

will be written to HDFS
Rather than written as intermediate data
One file per Mapper will be written

09-43

Practical Development Tips and Techniques

09-44

Hands-On Exercise
In this Hands-On Exercise you will write a Map-Only MapReduce
job using Counters
Please refer to the Hands-On Exercise Manual

09-45

Practical Development Tips and Techniques

09-46

Conclusion
In this chapter you have learned
Strategies for debugging your MapReduce code
How to write and subsequently view log files
How to retrieve job information with Counters
What issues to consider when using file compression
How to determine the optimal number of Reducers for a job
How to create Map-only MapReduce jobs

09-47

Chapter 10
More Advanced
MapReduce Programming

10-1

10-2

More Advanced MapReduce Programming

In this chapter you will learn
How to create custom Writables and WritableComparables
How to save binary data using SequenceFiles and Avro
How to build custom InputFormats and OutputFormats

10-3

More Advanced MapReduce Programming

A Recap of the MapReduce Flow
Custom Writables and WritableComparables
Saving Binary Data
Creating InputFormats and OutputFormats
Hands-On Exercise
Conclusion

10-4

Recap: Inputs to Mappers

10-5

Recap: Sort and Shuffle

10-6

Recap: Reducers to Outputs

10-7

More Advanced MapReduce Programming

A Recap of the MapReduce Flow
Custom Writables and WritableComparables
Saving Binary Data
Creating InputFormats and OutputFormats
Hands-On Exercise
Conclusion

10-8

Data Types in Hadoop

Writable

WritableComparable

IntWritable
LongWritable
Text

Defines a de/serialization
protocol. Every data type in
Hadoop is a Writable

Defines a sort order. All keys

must be WritableComparable

Concrete classes for different

data types

10-9

Box Classes in Hadoop

Hadoop data types are box classes
Text: string
IntWritable: int
LongWritable: long
FloatWritable: float

Writable defines wire transfer format

10-10

Creating a Complex Writable

Example: say we want a tuple (a, b)
We could artificially construct it by, for example, saying
Text t = new Text(a + "," + b);
...
String[] arr = t.toString().split(",");

Inelegant
Problematic
If a or b contained commas
Not always practical
Doesnt easily work for binary objects
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

10-11

The Writable Interface

public interface Writable {
void readFields(DataInput in);
void write(DataOutput out);
}

DataInput/DataOutput supports
boolean
byte, char (Unicode: 2 bytes)
double, float, int, long,
String (Unicode or UTF-8)
Line until line terminator
unsigned byte, short
byte array

10-12

Example: 3D Point Class

class Point3d implements Writable {
float x, y, z;
void readFields(DataInput in) {
x = in.readFloat();
y = in.readFloat();
z = in.readFloat();
}
void write(DataOutput out) {
out.writeFloat(x);
out.writeFloat(y);
out.writeFloat(z);
}
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

10-13

What About Binary Objects?

Solution: use byte arrays
Write idiom:
Serialize object to byte array
Write byte count
Write byte array
Read idiom:
Read byte count
Create byte array of proper size
Read byte array
Deserialize object

10-14

WritableComparable
WritableComparable is a sub-interface of Writable
Must implement compareTo, hashCode, equals methods
All keys in MapReduce must be WritableComparable

10-15

Example: 3D Point Class

class Point3d implements WritableComparable {
float x, y, z;
void readFields(DataInput in) {
x = in.readFloat();
y = in.readFloat();
z = in.readFloat();
}
void write(DataOutput out) {
out.writeFloat(x);
out.writeFloat(y);
out.writeFloat(z);
}
public int compareTo(Point3d p) {
// whatever ordering makes sense ... (e.g., closest to origin)
}
public boolean equals(Object o) {
Point3d p = (Point3d) o;
return this.x == p.x && this.y == p.y && this.z == p.z;
}
public int hashCode() {
// whatever makes sense ...
}
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

10-16

Comparators to Speed Up Sort and Shuffle

Recall that after each Map task, Hadoop must sort the keys
Keys are passed to a Reducer in sorted order
Nave approach: use the WritableComparables compareTo
method for each key
This requires deserializing each key, which could be time
consuming
Better approach: use a Comparator
A class which can compare two WritableComparables, ideally
just by looking at the two byte streams
Avoids the need for deserialization, if possible
Not always possible; depends on the actual key definition

10-17

Example Comparator for the Text Class

public class Text implements WritableComparable {
...
public static class Comparator extends WritableComparator {
public Comparator() {
super(Text.class);
}
public int compare(byte[] b1, int s1, int l1, byte[] b2, int s2, int l2)
{
int n1 = WritableUtils.decodeVIntSize(b1[s1]);
int n2 = WritableUtils.decodeVIntSize(b2[s2]);
return compareBytes(b1, s1+n1, l1-n1, b2, s2+n2, l2-n2);
}
}
...
static {
// register this comparator
WritableComparator.define(Text.class, new Comparator());
}
}

10-18

Using Custom Types in MapReduce Jobs

Use methods in JobConf to specify your custom key/value types
For output of Mappers:
conf.setMapOutputKeyClass()
conf.setMapOutputValueClass()

For output of Reducers:

conf.setOutputKeyClass()
conf.setOutputValueClass()

Input types are defined by InputFormat

See later

10-19

More Advanced MapReduce Programming

A Recap of the MapReduce Flow
Custom Writables and WritableComparables
Saving Binary Data
Creating InputFormats and OutputFormats
Hands-On Exercise
Conclusion

10-20

What Are SequenceFiles?

SequenceFiles are files containing binary-encoded key-value
pairs
Work naturally with Hadoop data types
SequenceFiles include metadata which identifies the data type
of the key and value
Actually, three file types in one
Uncompressed
Record-compressed
Block-compressed
Often used in MapReduce
Especially when the output of one job will be used as the input for
another
SequenceFileInputFormat
SequenceFileOutputFormat
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

10-21

Directly Accessing SequenceFiles

It is possible to directly access SequenceFiles from your code:
Configuration config = new Configuration();
SequenceFile.Reader reader =
new SequenceFile.Reader(FileSystem.get(config), path, config);
Text key = (Text) reader.getKeyClass().newInstance();
IntWritable value = (IntWritable) reader.getValueClass().newInstance();
while (reader.next(key, value)) {
// do something here
}
reader.close();

10-22

Problems With SequenceFiles

SequenceFiles are very useful but have some potential problems
They are only typically accessible via the Java API
Some work has been done to allow access from other languages
If the definition of the key or value object changes, the file
becomes unreadable

10-23

An Alternative to SequenceFiles: Avro

Apache Avro is a serialization format which is becoming a
popular alternative to SequenceFiles
Project was created by Doug Cutting, the creator of Hadoop
Self-describing file format
The schema for the data is included in the file itself
Compact file format
Portable across multiple languages
Support for C, C++, Java, Python, Ruby and others
Compatible with Hadoop
Via the AvroMapper and AvroReducer classes

10-24

More Advanced MapReduce Programming

A Recap of the MapReduce Flow
Custom Writables and WritableComparables
Saving Binary Data
Creating InputFormats and OutputFormats
Hands-On Exercise
Conclusion

10-25

Reprise: The Role of the InputFormat

10-26

Most Common InputFormats

Most common InputFormats:
TextInputFormat
KeyValueTextInputFormat
SequenceFileInputFormat
Others are available
NLineInputFormat
Every n lines of an input file is treated as a separate InputSplit
Configure in the driver code with
mapred.line.input.format.linespermap

MultiFileInputFormat
Abstract class which manages the use of multiple files in a
single task
You must supply a getRecordReader() implementation
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

10-27

How FileInputFormat Works

All file-based InputFormats inherit from FileInputFormat
FileInputFormat computes InputSplits based on the size of
each file, in bytes
HDFS block size is used as upper bound for InputSplit size
Lower bound can be specified in your driver code
Important: InputSplits do not respect record boundaries!

10-28

What RecordReaders Do
InputSplits are handed to the RecordReaders
Specified by the path, starting position offset, length
RecordReaders must:
Ensure each (key, value) pair is processed
Ensure no (key, value) pair is processed more than once
Handle (key, value) pairs which are split across InputSplits

10-29

Sample InputSplit

10-30

From InputSplits to RecordReaders

10-31

Writing Custom InputFormats

Use FileInputFormat as a starting point
Extend it
Write your own custom RecordReader
Override getRecordReader method in FileInputFormat
Override isSplittable if you dont want input files to be split

10-32

Reprise: Role of the OutputFormat

10-33

OutputFormat
OutputFormats work much like InputFormat classes
Custom OutputFormats must provide a RecordWriter
implementation

10-34

More Advanced MapReduce Programming

A Recap of the MapReduce Flow
Custom Writables and WritableComparables
Saving Binary Data
Creating InputFormats and OutputFormats
Hands-On Exercise
Conclusion

10-35

Hands-On Exercise
In this Hands-On Exercise you will create a custom Writable to
extend the word co-occurrence program you wrote earlier in the
course
Please refer to the Hands-On Exercise Manual

10-36

More Advanced MapReduce Programming

A Recap of the MapReduce Flow
Custom Writables and WritableComparables
Saving Binary Data
Creating InputFormats and OutputFormats
Hands-On Exercise
Conclusion

10-37

Conclusion
In this chapter you have learned
How to create custom Writables and WritableComparables
How to save binary data using SequenceFiles and Avro
How to build custom InputFormats and OutputFormats

10-38

Chapter 11
Joining Data Sets
in MapReduce Jobs

11-1

11-2

Joining Data Sets in MapReduce Jobs

In this chapter you will learn
How to write a Map-side join
How to implement the secondary sort
How to write a Reduce-side join

11-3

Joining Data Sets in MapReduce Jobs

Introduction
Map-Side Joins
The Secondary Sort
Reduce-Side Joins
Conclusion

11-4

Introduction
We frequently need to join data together from two sources as
part of a MapReduce job, such as
Lookup tables
Data from database tables
There are two fundamental approaches: Map-side joins and
Reduce-side joins
Map-side joins are easier to write, but have potential scaling
issues
We will investigate both types of joins in this chapter

11-5

But First
But first
Avoid writing joins in Java MapReduce if you can!
Abstractions such as Pig and Hive are much easier to use
Save hours of programming
If you are dealing with text-based data, there really is no reason
not to use Pig or Hive

11-6

Joining Data Sets in MapReduce Jobs

Introduction
Map-Side Joins
The Secondary Sort
Reduce-Side Joins
Conclusion

11-7

Map-Side Joins: The Algorithm

Basic idea for Map-side joins:
Load one set of data into memory, stored in an associative array
Key of the associative array is the join key
Map over the other set of data, and perform a lookup on the
associative array using the join key
If the join key is found, you have a successful join
Otherwise, do nothing

11-8

Map-Side Joins: Problems, Possible Solutions

Map-side joins have scalability issues
The associative array may become too large to fit in memory
Possible solution: break one data set into smaller pieces
Load each piece into memory individually, mapping over the
second data set each time
Then combine the result sets together

11-9

Joining Data Sets in MapReduce Jobs

Introduction
Map-Side Joins
The Secondary Sort
Reduce-Side Joins
Conclusion

11-10

Secondary Sort: Motivation

Recall that keys are passed to the Reducer in sorted order
The list of values for a particular key is not sorted
Order may well change between different runs of the MapReduce
job
Sometimes a job needs to receive the values for a particular key
in a sorted order
This is known as a secondary sort

11-11

Implementing the Secondary Sort

To implement a secondary sort, the intermediate key should be a
composite of the actual (natural) key and the value
Define a Partitioner which partitions just on the natural key
Define a Comparator class which sorts on the entire composite
key
Orders by natural key and, for the same natural key, on the value
portion of the key
Ensures that the keys are passed to the Reducer in the desired
order
Specified in the driver code by
conf.setOutputKeyComparatorClass(MyOKCC.class);

11-12

Implementing the Secondary Sort (contd)

Now we know that all values for the same natural key will go to
the same Reducer
And they will be in the order we desire
We must now ensure that all the values for the same natural key
are passed in one call to the Reducer
Achieved by defining a Grouping Comparator class which
partitions just on the natural key
Determines which keys and values are passed in a single call to
the Reducer.
Specified in the driver code by
conf.setOutputValueGroupingComparator(MyOVGC.class);

11-13

Secondary Sort: Example

Assume we have input with (key, value) pairs like this
foo
foo
bar
baz
foo
bar
baz

98
101
12
18
22
55
123

We want the Reducer to receive the intermediate data for each

key in descending numerical order

11-14

Secondary Sort: Example (contd)

Write the Mapper such that they key is a composite of the natural
key and value
For example, intermediate output may look like this:
('foo#98', 98)
('foo#101', 101)
('bar#12',12)
('baz#18', 18)
('foo#22', 22)
('bar#55', 55)
('baz#123', 123)

11-15

Secondary Sort: Example (contd)

Write a class that extends WritableComparator and sorts on
natural key, and for identical natural keys, sorts on the value
portion in descending order
Just override compare(WritableComparable,
WritableComparable)
Supply a reference to this class in your driver using the
JobConf.setOutputKeyComparatorClass method
Will result in keys being passed to the Reducer in this order:
('bar#55', 55)
('bar#12', 12)
('baz#123', 123)
('baz#18', 18)
('foo#101', 101)
('foo#98', 98)
('foo#22', 22)
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

11-16

Secondary Sort: Example (contd)

Finally, write another WritableComparator subclass which just
examines the first portion of the key
Again, just override compare(WritableComparable,
WritableComparable)
Supply a reference to this class in your driver using the
JobConf.setOutputValueGroupingComparator method
This will ensure that values associated with the same natural key
will be sent to the same pass of the Reducer
But theyre sorted in descending order, as we required

11-17

Joining Data Sets in MapReduce Jobs

Introduction
Map-Side Joins
The Secondary Sort
Reduce-Side Joins
Conclusion

11-18

Reduce-Side Joins: The Basic Concept

For a Reduce-side join, the basic concept is:
Map over both data sets
Emit a (key, value) pair for each record
Key is the join key, value is the entire record
In the Reducer, do the actual join
Because of the Shuffle and Sort, values with the same key
are brought together

11-19

Reduce-Side Joins: Example

Example input data:
EMP: 42, Aaron, loc(13)
LOC: 13, New York City

Required output:
EMP: 42, Aaron, loc(13), New York City

11-20

Example Record Data Structure

A data structure to hold a record could look like this:
class Record {
enum Typ { emp, loc };
Typ type;

String empName;
int empId;
int locId;
String locationName;

11-21

Reduce-Side Join: Mapper

void map(k, v) {
Record r = parse(v);
emit (r.locId, r);
}

11-22

Reduce-Side Join: Reducer

void reduce(k, values) {
Record thisLocation;
List<Record> employees;
for (Record v in values) {
if (v.type == Typ.loc) {
thisLocation = v;
} else {
employees.add(v);
}
}
for (Record e in employees) {
e.locationName = thisLocation.locationName;
emit(e);
}
}
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

11-23

Scalability Problems With Our Reducer

All employees for a given location must potentially be buffered in
the Reducer
Could result in out-of-memory errors for large data sets
Solution: Ensure the location record is the first one to arrive at
the Reducer
Using a Secondary Sort

11-24

A Better Intermediate Key

class LocKey {
boolean isPrimary;
int locId;
public int compareTo(LocKey k) {
if (locId == k.locId) {
return Boolean.compare(k.isPrimary, isPrimary);
} else {
return Integer.compare(locId, k.locId);
}
}
public int hashCode() {
return locId;
}
}

11-25

A Better Intermediate Key (contd)

11-26

A Better Intermediate Key (contd)

The compareTo method ensures that primary keys will

publicsort
intearlier
hashCode()
{
than non-primary
keys for the same location
return locId;

}
}

11-27

A Better Mapper
void map(k, v) {
Record r = parse(v);
if (r.type == Typ.emp) {
emit (FK(r.locId), r);
} else {
emit (PK(r.locId), r);
}
}

11-28

A Better Reducer
Record thisLoc;
void reduce(k, values) {
for (Record v in values) {
if (v.type == Typ.loc) {
thisLoc = v;
} else {
v.locationName = thisLoc.locationName;
emit(v);
}
}
}

11-29

Create a Grouping Comparator

class LocIDComparator extends WritableComparator {
public int compare(Record r1, Record r2) {
return Integer.compare(r1.locId, r2.locId);
}
}

11-30

And Configure Hadoop To Use It In The Driver

conf.setOutputValueGroupingComparator(LocIDComparator.class)

11-31

Joining Data Sets in MapReduce Jobs

Introduction
Map-Side Joins
The Secondary Sort
Reduce-Side Joins
Conclusion

11-32

Conclusion
In this chapter you have learned
How to join write a Map-side join
How to implement the secondary sort
How to write a Reduce-side join

11-33

Chapter 12
Graph Manipulation
in MapReduce

12-1

12-2

Graph Manipulation in MapReduce

In this chapter you will learn
Best practices for representing graphs in Hadoop
How to implement a single source shortest path algorithm in
MapReduce

12-3

Graph Manipulation in MapReduce

Introduction
Representing Graphs
Implementing Single Source Shortest Path
Conclusion

12-4

Introduction: What Is A Graph?

Loosely speaking, a graph is a set of vertices, or nodes,
connected by edges, or lines
There are many different types of graphs
Directed
Undirected
Cyclic
Acyclic
Weighted
Unweighted
DAG (Directed, Acyclic Graph) is a very common graph type

12-5

What Can Graphs Represent?

Graphs are everywhere
Hyperlink structure of the Web
Physical structure of computers on a network
Roadmaps
Airline flights
Social networks

12-6

Examples of Graph Problems

Finding the shortest path through a graph
Routing Internet traffic
Giving driving directions
Finding the minimum spanning tree
Lowest-cost way of connecting all nodes in a graph
Example: telecoms company laying fiber
Must cover all customers
Need to minimize fiber used
Finding maximum flow
Move the most amount of traffic through a network
Example: airline scheduling

12-7

Examples of Graph Problems (contd)

Finding critical nodes without which a graph would break into
disjoint components
Controlling the spread of epidemics
Breaking up terrorist cells

12-8

Graphs and MapReduce

Graph algorithms typically involve:
Performing computations at each vertex
Traversing the graph in some manner
Key questions:
How do we represent graph data in MapReduce?
How do we traverse a graph in MapReduce?

12-9

Graph Manipulation in MapReduce

Introduction
Representing Graphs
Implementing Single Source Shortest Path
Conclusion

12-10

Representing Graphs
Imagine we want to represent this simple graph:
Two approaches:
Adjacency matrices
Adjacency lists

2"
1"
3"

12-11

Adjacency Matrices
Represent the graph as an n x n square matrix

v1 v2 v3 v4
v1 0

v2 1

v3 1

v4 1

1"
3"

12-12

Adjacency Matrices: Critique

Advantages:
Naturally encapsulates iteration over nodes
Rows and columns correspond to inlinks and outlinks
Disadvantages:
Lots of zeros for sparse matrices
Lots of wasted space

12-13

Adjacency Lists
Take an adjacency matrix and throw away all the zeros

v1 v2 v3 v4
v1 0

v2 1

v3 1

v4 1

v1: v2, v4
v2: v1, v3, v4
v3: v1
v4: v1, v3

12-14

Adjacency Lists: Critique

Advantages:
Much more compact representation
Easy to compute outlinks
Graph structure can be broken up and distributed
Disadvantages:
More difficult to compute inlinks

12-15

Encoding Adjacency Lists

Adjacency lists are the preferred way of representing graphs in
MapReduce
Typically we represent each vertex (node) with an id number
A field of type long usually suffices
Typical encoding format (Writable)
long: vertex id of the source
int: number of outgoing edges
Sequence of longs: destination vertices

v1: v2, v4
v2: v1, v3, v4
v3: v1
v4: v1, v3

1: [2] 2, 4
2: [3] 1, 3, 4
3: [1] 1
4: [2] 1, 3

12-16

Graph Manipulation in MapReduce

Introduction
Representing Graphs
Implementing Single Source Shortest Path
Conclusion

12-17

Single Source Shortest Path

Problem: find the shortest path from a source node to one or
more target nodes
Serial algorithm: Dijkstras Algorithm
Not suitable for parallelization
MapReduce algorithm: parallel breadth-first search

12-18

Parallel Breadth-First Search

The algorithm, intuitively:
Distance to the source = 0
For all nodes directly reachable from the source, distance = 1
For all nodes reachable from some node n in the graph, distance
from source = 1 + min(distance to that node)

12-19

Parallel Breadth-First Search: Algorithm

Mapper:
Input key is some vertex id
Input value is D (distance from source), adjacency list
Processing: For all nodes in the adjacency list, emit (node id, D +
1)
If the distance to this node is D, then the distance to any node
reachable from this node is D + 1
Reducer:
Receives vertex and list of distance values
Processing: Selects the shortest distance value for that node

12-20

Iterations of Parallel BFS

A MapReduce job corresponds to one iteration of parallel
breadth-first search
Each iteration advances the known frontier by one hop
Iteration is accomplished by using the output from one job as the
input to the next
How many iterations are needed?
Multiple iterations are needed to explore the entire graph
As many as the diameter of the graph
Graph diameters are surprisingly small, even for large graphs
Six degrees of separation
Controlling iterations in Hadoop
Use counters; when you reach a node, count it
At the end of each iteration, check the counters
When youve reached all the nodes, youre finished
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

12-21

One More Trick: Preserving Graph Structure

Characters of Parallel BFS
Mappers emit distances, Reducers select the shortest distance
Output of the Reducers becomes the input of the Mappers for the
next iteration
Problem: where did the graph structure (adjacency lists) go?
Solution: Mapper must emit the adjacency lists as well
Mapper emits two types of key-value pairs
Representing distances
Representing adjacency lists
Reducer recovers the adjacency list and preserves it for the next
iteration

12-22

Parallel BFS: Pseudo-Code

12-23

Parallel BFS: Demonstration

Your instructor will now demonstrate the parallel breadth-first
search algorithm

12-24

Graph Algorithms: General Thoughts

MapReduce is adept at manipulating graphs
Store graphs as adjacency lists
Typically, MapReduce graph algorithms are iterative
Iterate until some termination condition is met
Remember to pass the graph structure from one iteration to the
next

12-25

Graph Manipulation in MapReduce

Introduction
Representing Graphs
Implementing Single Source Shortest Path
Conclusion

12-26

Conclusion
In this chapter you have learned
Best practices for representing graphs in Hadoop
How to implement a single source shortest path algorithm in
MapReduce

12-27

Chapter 13
An Introduction
To Oozie

13-1

13-2

An Introduction to Oozie
In this chapter you will learn
What Oozie is
How to create Oozie workflows

13-3

An Introduction to Oozie
The Motivation for Oozie
Creating Oozie workflows
Hands-On Exercise
Conclusion

13-4

The Motivation for Oozie

Many problems cannot be solved with a single MapReduce job
Instead, a workflow of jobs must be created
Simple workflow:
Run JobA
Use output of JobA as input to JobB
Use output of JobB as input to JobC
Output of JobC is the final required output
Easy if the workflow is linear like this
Can be created as standard Driver code

13-5

The Motivation for Oozie (contd)

If the workflow is more complex, Driver code becomes much
more difficult to maintain
Example: running multiple jobs in parallel, using the output from
all of those jobs as the input to the next job
Example: including Hive or Pig jobs as part of the workflow

13-6

What is Oozie?
Oozie is a workflow engine
Runs on a server
Typically outside the cluster
Runs workflows of Hadoop jobs
Including Pig, Hive, Sqoop jobs
Submits those jobs to the cluster based on a workflow definition
Workflow definitions are submitted via HTTP
Jobs can be run at specific times
One-off or recurring jobs
Jobs can be run when data is present in a directory

13-7

An Introduction to Oozie
The Motivation for Oozie
Creating Oozie workflows
Hands-On Exercise
Conclusion

13-8

Oozie Workflow Basics

Oozie workflows are written in XML
Workflow is a collection of actions
MapReduce jobs, Pig jobs, Hive jobs etc.
A workflow consists of control flow nodes and action nodes
Control flow nodes define the beginning and end of a workflow
They provide methods to determine the workflow execution path
Example: Run multiple jobs simultaneously
Action nodes trigger the execution of a processing task, such as
A MapReduce job
A Pig job
A Sqoop data import job

13-9

Simple Oozie Example

Simple example workflow for WordCount:

13-10

Simple Oozie Example (contd)

<workflow-app name='wordcount-wf' xmlns="uri:oozie:workflow:0.1">
<start to='wordcount'/>
<action name='wordcount'>
<map-reduce>
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<configuration>
<property>
<name>mapred.mapper.class</name>
<value>org.myorg.WordCount.Map</value>
</property>
<property>
<name>mapred.reducer.class</name>
<value>org.myorg.WordCount.Reduce</value>
</property>
<property>
<name>mapred.input.dir</name>
<value>${inputDir}</value>
</property>
<property>
<name>mapred.output.dir</name>
<value>${outputDir}</value>
</property>
</configuration>
</map-reduce>
<ok to='end'/>
<error to='end'/>
</action>
<kill name='kill'>
<message>Something went wrong: ${wf:errorCode('wordcount')}</message>
</kill/>
<end name='end'/>
</workflow-app>
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

13-11

Simple Oozie Example (contd)

A workflow is wrapped in the workflow entity

13-12

Simple Oozie Example (contd)

The start node is the control node which tells

Oozie which workflow node should be run first.
There must be one start node in an Oozie
workflow. In our example, we are telling Oozie to
start by transitioning to the wordcount workflow
node.

13-13

Simple Oozie Example (contd)

The wordcount action node defines a mapreduce action a standard Java MapReduce
job.

13-14

Simple Oozie Example (contd)

Within the action, we define the jobs properties.

13-15

Simple Oozie Example (contd)

We specify what to do if the action ends

successfully, and what to do if it fails. In this
example, in both cases we transition to the end
node.

13-16

Simple Oozie Example (contd)

Every workflow must have an end node. This

indicates that the workflow has completed
successfully.

13-17

Simple Oozie Example (contd)

If the workflow reaches a kill node, it will kill all

running actions and then terminate with an error.
A workflow can have zero or more kill nodes.

13-18

Other Oozie Control Nodes

A decision control node allows Oozie to determine the workflow
execution path based on some criteria
Similar to a switch-case statement
fork and join control nodes split one execution path into
multiple execution paths which run concurrently
fork splits the execution path
join waits for all concurrent execution paths to complete before
proceeding
fork and join are used in pairs

13-19

Oozie Workflow Action Nodes

Node%Name%

Descrip.on%

map-reduce

Runs%either%a%Java%MapReduce%or%Streaming%job%

Create%directories,%move%or%delete%les%or%directories%

java

Runs%the%main()%method%in%the%specied%Java%class%as%a%single<Map,%
Map<only%job%on%the%cluster%

pig

Runs%a%Pig%job%

hive

Runs%a%Hive%job%

sqoop

Runs%a%Sqoop%job%

Sends%an%e<mail%message%

13-20

Submitting an Oozie Workflow

To submit an Oozie workflow using the command-line tool:
$ oozie job oozie http://<oozie_server>/oozie
config config_file -run

Oozie can also be called from within a Java program

Via the Oozie client API

13-21

More on Oozie
For full documentation on Oozie, refer to
http://docs.cloudera.com/

13-22

An Introduction to Oozie
The Motivation for Oozie
Creating Oozie workflows
Hands-On Exercise
Conclusion

13-23

Hands-On Exercise
In this Hands-On Exercise you will run Oozie jobs
Please refer to the Hands-On Exercise Manual

13-24

An Introduction to Oozie
The Motivation for Oozie
Creating Oozie workflows
Hands-On Exercise
Conclusion

13-25

Conclusion
In this chapter you have learned
What Oozie is
How to create Oozie workflows

13-26

Chapter 14
Conclusion

14-1

Conclusion
During this course, you have learned:
The core technologies of Hadoop
How HDFS and MapReduce work
What other projects exist in the Hadoop ecosystem
How to develop MapReduce jobs
How Hadoop integrates into the datacenter
Algorithms for common MapReduce tasks
How to create large workflows using Oozie

14-2

Conclusion (contd)
How Hive and Pig can be used for rapid application development
Best practices for developing and debugging MapReduce jobs
Advanced features of the Hadoop API
How to handle graph manipulation problems with MapReduce

14-3

Certification
You can now take the Cloudera Certified Developer for Apache
Hadoop exam
Your instructor will give you information on how to access the
exam

Thank you for attending the course!

If you have any questions or comments, please contact us via
http://www.cloudera.com/
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

14-4

Appendix A:
Cloudera Enterprise

A-1

Cloudera Enterprise
Reduces the risks of running Hadoop in production
Improves consistency, compliance, and administrative overhead
Management Suite components:
- Cloudera Manager (CM)
- Activity Monitor
- Resource Manager
- Authorization Manager
- The Flume User Interface

24x7 Production support for CDH and certified integrations

A-2

Cloudera Manager
Cloudera Manager (CM) is designed to make installing and
managing your cluster very easy
View a dashboard of cluster status
Modify configuration parameters
Easily start and stop master and slave daemons
Easily retrieve the configuration files required for client
machines to access the cluster

A-3

Cloudera Manager (contd)

A-4

Activity Monitor
Activity Monitor gives an in-depth, comprehensive view of what
is happening on the cluster, in real-time
And what has happened in the past
Compare the performance of similar jobs
Store historical data
Chart metrics on cluster performance

A-5

Activity Monitor (contd)

A-6

Reporting
Cloudera Manager can report on resource usage in the cluster
Helpful for auditing and internal billing

A-7

Authorization Manager
Authorization Manager allows you to provision users, and
manage user activity, within the cluster

A-8

Conclusion
Cloudera Enterprise makes it easy to run open source Hadoop in
production
Includes
Cloudera Distribution including Apache Hadoop (CDH)
Cloudera Management Suite (CMS)
Production Support
Cloudera Management Suite enables you to:
Simplify and accelerate Hadoop deployment
Reduce the costs and risks of adopting Hadoop in production
Reliably operate Hadoop in production with repeatable success
Apply SLAs to Hadoop
Increase control over Hadoop cluster provisioning and
management
Copyright 2010-2012 Cloudera. All rights reserved. Not to be reproduced without prior written consent.

A-9

The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life
From Everand
The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life
Mark Manson
4/5 (6458)
Principles: Life and Work
From Everand
Principles: Life and Work
Ray Dalio
4/5 (643)
The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are
From Everand
The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are
Brené Brown
4/5 (1175)
Never Split the Difference: Negotiating As If Your Life Depended On It
From Everand
Never Split the Difference: Negotiating As If Your Life Depended On It
Chris Voss
4.5/5 (1005)
The Glass Castle: A Memoir
From Everand
The Glass Castle: A Memoir
Jeannette Walls
4.5/5 (1856)
Grit: The Power of Passion and Perseverance
From Everand
Grit: The Power of Passion and Perseverance
Angela Duckworth
4/5 (650)
The Perks of Being a Wallflower
From Everand
The Perks of Being a Wallflower
Stephen Chbosky
4.5/5 (4103)
Sing, Unburied, Sing: A Novel
From Everand
Sing, Unburied, Sing: A Novel
Jesmyn Ward
4/5 (1267)
Her Body and Other Parties: Stories
From Everand
Her Body and Other Parties: Stories
Carmen Maria Machado
4/5 (903)
Shoe Dog: A Memoir by the Creator of Nike
From Everand
Shoe Dog: A Memoir by the Creator of Nike
Phil Knight
4.5/5 (629)
Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race
From Everand
Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race
Margot Lee Shetterly
4/5 (1022)
The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers
From Everand
The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers
Ben Horowitz
4.5/5 (361)
Steve Jobs
From Everand
Steve Jobs
Walter Isaacson
4.5/5 (1139)
Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future
From Everand
Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future
Ashlee Vance
4.5/5 (582)
The Emperor of All Maladies: A Biography of Cancer
From Everand
The Emperor of All Maladies: A Biography of Cancer
Siddhartha Mukherjee
4.5/5 (298)
A Man Called Ove: A Novel
From Everand
A Man Called Ove: A Novel
Fredrik Backman
4.5/5 (5181)
The Yellow House: A Memoir (2019 National Book Award Winner)
From Everand
The Yellow House: A Memoir (2019 National Book Award Winner)
Sarah M. Broom
4/5 (100)
Angela's Ashes: A Memoir
From Everand
Angela's Ashes: A Memoir
Frank McCourt
4.5/5 (943)
Brooklyn: A Novel
From Everand
Brooklyn: A Novel
Colm Toibin
3.5/5 (2141)
The Little Book of Hygge: Danish Secrets to Happy Living
From Everand
The Little Book of Hygge: Danish Secrets to Happy Living
Meik Wiking
3.5/5 (464)
The World Is Flat 3.0: A Brief History of the Twenty-first Century
From Everand
The World Is Flat 3.0: A Brief History of the Twenty-first Century
Thomas L. Friedman
3.5/5 (2289)
Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America
From Everand
Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America
Gilbert King
4.5/5 (280)
Yes Please
From Everand
Yes Please
Amy Poehler
4/5 (2016)
Bad Feminist: Essays
From Everand
Bad Feminist: Essays
Roxane Gay
4/5 (1090)
The Art of Racing in the Rain: A Novel
From Everand
The Art of Racing in the Rain: A Novel
Garth Stein
4/5 (4372)
A Tree Grows in Brooklyn
From Everand
A Tree Grows in Brooklyn
Betty Smith
4.5/5 (2033)
The Woman in Cabin 10
From Everand
The Woman in Cabin 10
Ruth Ware
3.5/5 (2814)
The Outsider: A Novel
From Everand
The Outsider: A Novel
Stephen King
4/5 (2885)
A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story
From Everand
A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story
Dave Eggers
3.5/5 (233)
The Sympathizer: A Novel (Pulitzer Prize for Fiction)
From Everand
The Sympathizer: A Novel (Pulitzer Prize for Fiction)
Viet Thanh Nguyen
4.5/5 (141)
Team of Rivals: The Political Genius of Abraham Lincoln
From Everand
Team of Rivals: The Political Genius of Abraham Lincoln
Doris Kearns Goodwin
4.5/5 (244)
Wolf Hall: A Novel
From Everand
Wolf Hall: A Novel
Hilary Mantel
4/5 (4135)
On Fire: The (Burning) Case for a Green New Deal
From Everand
On Fire: The (Burning) Case for a Green New Deal
Naomi Klein
4/5 (78)
Rise of ISIS: A Threat We Can't Ignore
From Everand
Rise of ISIS: A Threat We Can't Ignore
Jay Sekulow
3.5/5 (144)
Manhattan Beach: A Novel
From Everand
Manhattan Beach: A Novel
Jennifer Egan
3.5/5 (919)
Fear: Trump in the White House
From Everand
Fear: Trump in the White House
Bob Woodward
3.5/5 (836)
John Adams
From Everand
John Adams
David McCullough
4.5/5 (2546)
Artistic Thinking As Transcognitive Practice
No ratings yet
Artistic Thinking As Transcognitive Practice
12 pages
The Unwinding: An Inner History of the New America
From Everand
The Unwinding: An Inner History of the New America
George Packer
4/5 (45)
The Light Between Oceans: A Novel
From Everand
The Light Between Oceans: A Novel
M.L. Stedman
4.5/5 (815)
Little Women
From Everand
Little Women
Louisa May Alcott
4.5/5 (2369)
The Constant Gardener: A Novel
From Everand
The Constant Gardener: A Novel
John le Carré
4/5 (278)
Lea History Confession Indulgences 1896
No ratings yet
Lea History Confession Indulgences 1896
546 pages
MP 4 JSD
No ratings yet
MP 4 JSD
8 pages
Holiday Homework Xii History
No ratings yet
Holiday Homework Xii History
3 pages
Mac 221
No ratings yet
Mac 221
169 pages
Unit7 Speaking
No ratings yet
Unit7 Speaking
5 pages
Unit 5 Ielts Masterclass
No ratings yet
Unit 5 Ielts Masterclass
5 pages
Unit 4 Unit Wise Question
No ratings yet
Unit 4 Unit Wise Question
20 pages
Fourier Series: Applied Mathematics - III
No ratings yet
Fourier Series: Applied Mathematics - III
78 pages
Lesson Plan in Science K1
No ratings yet
Lesson Plan in Science K1
40 pages
Wisdom Diary On Ministry
100% (2)
Wisdom Diary On Ministry
15 pages
PsychologicalEssentialism EJSP2001
No ratings yet
PsychologicalEssentialism EJSP2001
20 pages
02.RBS 6000 Config PDF
No ratings yet
02.RBS 6000 Config PDF
66 pages
A Parable of God's Love For Sinners - Luke 15
No ratings yet
A Parable of God's Love For Sinners - Luke 15
15 pages
Jha BrahmanicalIntoleranceEarly 2016
No ratings yet
Jha BrahmanicalIntoleranceEarly 2016
9 pages
CH 1 To CH 3 10th Class
No ratings yet
CH 1 To CH 3 10th Class
9 pages
Anna Karenina by Leo Tolstoy
No ratings yet
Anna Karenina by Leo Tolstoy
2 pages
Jan 23 2022
No ratings yet
Jan 23 2022
9 pages
LearnEnglish Speaking B1 Agreeing and Disagreeing
No ratings yet
LearnEnglish Speaking B1 Agreeing and Disagreeing
2 pages
Mit Ocw: 18.703 Modern Algebra Prof. James Mckernan
No ratings yet
Mit Ocw: 18.703 Modern Algebra Prof. James Mckernan
10 pages
Rupa Goswami - Navadvipa Stuti (Ingles)
No ratings yet
Rupa Goswami - Navadvipa Stuti (Ingles)
4 pages
Ontological Engineering: Delivered by Joel Anandraj.E Ap/It
No ratings yet
Ontological Engineering: Delivered by Joel Anandraj.E Ap/It
39 pages
Superiority of Christianity Over Other Religions On Earth by Pastor Paul Rika Ebook
100% (1)
Superiority of Christianity Over Other Religions On Earth by Pastor Paul Rika Ebook
76 pages
Chapter 8
No ratings yet
Chapter 8
5 pages
Lesson 2 - Teachers Guide
No ratings yet
Lesson 2 - Teachers Guide
8 pages
LP Spaces For P in 01
No ratings yet
LP Spaces For P in 01
4 pages
Identify The Choice That Best Completes The Statement or Answers The Question. Bcde Opqr DE PQ OR OP QR MNO PQR, MN PR M P NO QR N Q ABC
No ratings yet
Identify The Choice That Best Completes The Statement or Answers The Question. Bcde Opqr DE PQ OR OP QR MNO PQR, MN PR M P NO QR N Q ABC
32 pages
PlateMaker Advanced Tutorial Sample
No ratings yet
PlateMaker Advanced Tutorial Sample
8 pages
Netxms Admin
No ratings yet
Netxms Admin
546 pages
Final Evaluation Rose
No ratings yet
Final Evaluation Rose
7 pages