A standalone MATLAB mapreduce application is a self-contained application encompassing all the necessary components for independent execution against a Hadoop cluster. It contains a:
mapper function (one or more) written in MATLAB.
reducer function written in MATLAB.
MATLAB application script or function invoking the mapper and reducer functions.
All components are packaged into a single standalone application to be
executed against a Hadoop cluster. Standalone applications can be packaged using the
Application Compiler app or from the MATLAB command prompt using the mcc
command.
For more information, see Workflow to Run Compiled Standalone Applications Against a Hadoop Cluster.
deploytool | Open a list of application deployment apps |
mcc | Compile MATLAB functions for deployment |
mapreduce | Programming technique for analyzing data sets that do not fit in memory |
mapreducer | Define deployed execution for mapreduce |
matlab.mapreduce.DeployHadoopMapReducer | Configure a MapReduce application for deployment against Hadoop |
Workflow to Run Compiled Standalone Applications Against a Hadoop Cluster
Review workflow on how to create a compiled standalone application and run it against a Hadoop cluster.
Example on Running a Standalone MATLAB MapReduce Application
Try an example on creating a standalone MATLAB MapReduce application using the mcc
command, and run it against a Hadoop cluster.