stuartsierra / Clojure Hadoop

UP-TO-DATE fork with more recent maintenance is here: https://github.com/alexott/clojure-hadoop

clojure-hadoop

A library to assist in writing Hadoop MapReduce jobs in Clojure.

by Stuart Sierra http://stuartsierra.com/

For stable releases, see http://stuartsierra.com/software/clojure-hadoop

For more information on Clojure, http://clojure.org/ on Hadoop, http://hadoop.apache.org/

Also see my presentation about this library at http://vimeo.com/7669741

Copyright (c) Stuart Sierra, 2009. All rights reserved. The use and distribution terms for this software are covered by the Eclipse Public License 1.0 (http://opensource.org/licenses/eclipse-1.0.php) which can be found in the file LICENSE.html at the root of this distribution. By using this software in any fashion, you are agreeing to be bound by the terms of this license. You must not remove this notice, or any other, from this software.

DEPENDENCIES

This library requires Java 6 JDK, http://java.sun.com/

Building from source requires Apache Maven 2, http://maven.apache.org/

BUILDING

If you downloaded the library distribution as a .zip or .tar file, everything is pre-built and there is nothing you need to do.

If you downloaded the sources from Git, then you need to run the build with Maven. In the top-level directory of this project, run:

mvn assembly:assembly

This compiles and builds the JAR files.

You can find these files in the "target" directory (replace ${VERSION} with the current version number of this library):

clojure-hadoop-${VERSION}-examples.jar :

    This JAR contains all dependencies, including all of Hadoop
    0.18.3.  You can use this JAR to run the example MapReduce
    jobs from the command line.  This file is ONLY for running the
    examples.


clojure-hadoop-${VERSION}-job.jar :

    This JAR contains the clojure-hadoop libraries and Clojure
    1.0.  It is suitable for inclusion in the "lib" directory of a
    JAR file submitted as a Hadoop job.


clojure-hadoop-${VERSION}.jar :

    This JAR contains ONLY the clojure-hadoop libraries.  It can
    be placed in the "lib" directory of a JAR file submitted as a
    Hadoop job; that JAR must also include the Clojure 1.0 JAR.

RUNNING THE EXAMPLES

After building, copy the file from

target/clojure-hadoop-${VERSION}-examples.jar

to something short, like "examples.jar". Each of the *.clj files in the src/examples directory contains instructions for running that example.

USING THE LIBRARY IN HADOOP

After building, include the "clojure-hadoop-${VERSION}-job.jar" file in the lib/ directory of the JAR you submit as your Hadoop job.

DEPENDING ON THE LIBRARY WITH MAVEN

You can depend on clojure-hadoop in your Maven 2 projects by adding the following lines to your pom.xml:

<dependencies>
  ...

  <dependency>
    <groupId>com.stuartsierra</groupId>
    <artifactId>clojure-hadoop</artifactId>
    <version>${VERSION}</version>
  </dependency>

  ...
</dependencies>
...
<repositories>
  ...
  <!-- For released versions: -->
  <repository>
    <id>stuartsierra-releases</id>
    <name>Stuart Sierra's personal Maven 2 release repository</name>
    <url>http://stuartsierra.com/maven2</url>
  </repository>

  <!-- For SNAPSHOT versions: -->
  <repository>
    <id>stuartsierra-snapshots</id>
    <name>Stuart Sierra's personal Maven 2 SNAPSHOT repository</name>
    <url>http://stuartsierra.com/m2snapshots</url>
  </repository>
  ...
</repositories>

USING THE LIBRARY

This library provides different layers of abstraction away from the raw Hadoop API.

Layer 1: clojure-hadoop.imports

Provides convenience functions for importing the many classes and
interfaces in the Hadoop API.

Layer 2: clojure-hadoop.gen

Provides gen-class macros to generate the multiple classes needed
for a MapReduce job.  See the example file "wordcount1.clj" for a
demonstration of these macros.

Layer 3: clojure-hadoop.wrap

clojure-hadoop.wrap: provides wrapper functions that automatically
convert between Hadoop Text objects and Clojure data structures.
See the example file "wordcount2.clj" for a demonstration of these
wrappers.

Layer 4: clojure-hadoop.job

Provides a complete implementation of a Hadoop MapReduce job that
can be dynamically configured to use any Clojure functions in the
map and reduce phases.  See the example file "wordcount3.clj" for
a demonstration of this usage.

Layer 5: clojure-hadoop.defjob

A convenient macro to configure MapReduce jobs with Clojure code.
See the example files "wordcount4.clj" and "wordcount5.clj" for
demonstrations of this macro.