Documentation

Hadoop Integration

Deploy MATLAB® code against Hadoop®

There are two ways to deploy MapReduce applications. You can create a deployable archive or you can create a standalone application to run against Hadoop. Create a deployable archive if you want to use a Hadoop scheduling framework or if you want to integrate MATLAB code with an existing Hadoop job flow. A deployable archive includes a datastore, a map function, and a reduce function. A standalone application includes MATLAB code that uses a datastore, a mapreducer, and a mapreduce function. The execution of mapreduce function generates a Hadoop command that submits a job to Hadoop.

Deployable archive and standalone applications that run against Hadoop are supported only on Linux®.

Install MATLAB Runtime to deploy MapReduce applications. For more information, see Download and Install the MATLAB Runtime.

Apps

Hadoop Compiler Package MATLAB programs for deployment to Hadoop clusters as MapReduce programs

Functions

mapreduce Programming technique for analyzing data sets that do not fit in memory
mapreducer Define deployed execution for mapreduce
deploytool Compile and package functions for external deployment
hadoopCompiler Build and package MapReduce applications for deployment against Hadoop
mcc Compile MATLAB functions for deployment

Classes

matlab.mapreduce.DeployHadoopMapReducer Configure a MapReduce application for deployment against Hadoop

Examples and How To

Concepts

Related Information

Was this topic helpful?
MATLAB Compiler Documentation
Other Documentation
Support