mapreduceAssuming you have installed Hadoop on your cluster, if not please follow

This is the machine config of my cluster nodes, though the steps that follow could be followed with your installation/machine configs

pkommireddi@pkommireddi-wsl:/tools/hadoop/pig-0.9.1/lib$ uname -a

Linux pkommireddi-wsl 2.6.32-37-generic #81-Ubuntu SMP Fri Dec 2 20:32:42 UTC 2011 x86_64 GNU/Linux

Pig requires that the snappy jar and native be available on its classpath when a script is run.

The pig client here is installed at /tools/hadoop and the jar needs to be placed within $PIG_HOME/lib.


Also, you need to point PIG to the snappy native

export PIG_OPTS=”$PIG_OPTS -Djava.library.path=$HADOOP_HOME/lib/native/Linux-amd64-64″

hadoop pigNow you have 2 ways to use map output compression in the Pig scripts:

  1. Follow instructions on to set map output compression at a cluster level
  2. Use Pig’s “set” keyword for per job level configuration
  1. set true;


This should get you going with using Snappy for Map output compression with Pig. You can read and write Snappy compressed files as well, though I would not recommend doing that as its not very efficient space-wise compared to other compression algorithms. There is work being done to be able to use Snappy for creating intermediate/temporary files between multiple MR jobs. You can watch the work item here

Using Snappy for Native Java MapReduce:

Set Configuration parameters for Map output compression

Configuration conf = new Configuration();

conf.setBoolean(“”, true);


Set Configuration parameters for Snappy compressed intermediate Sequence Files


SequenceFileOutputFormat.setOutputCompressionType(conf, CompressionType.BLOCK); //Block level is better than Record level, in most cases

SequenceFileOutputFormat.setCompressOutput(conf, true);



  1. Map tasks begin transferring data sooner compared to Gzip or Bzip (though more data needs to be transferred to Reduce tasks)
  2. Reduce tasks run faster with better decompression speeds
  3. Snappy is not CPU intensive – which means MR tasks have more CPU for user operations

What you SHOULD use Snappy for

Map output: Snappy works great if you have large amounts of data flowing from Mappers to the Reducers (you might not see a significant difference if data volume between Map and Reduce is low)

Temporary Intermediate files (not available currently as of Pig 0.9.2, applicable only to native Map Reduce) : If you have a series of MR jobs chained together, Snappy compression is a good way to store the intermediate files. Please do make sure these intermediate files are cleaned up soon enough so we don’t have disk space issues on the cluster.

What you should NOT use Snappy for

Permanent Storage: Snappy compression is not efficient space-wise and it is expensive to store data on HDFS (3-way replication)

Plain text files: Like Gzip, Snappy is not splittable. Do not store plain text files in Snappy compressed form, instead use a container like SequenceFile. Source


Subscribe to our Newsletter

Stay up-to-date with the latest big data news.