The Umbrella Security Labs research team is constantly processing terabytes of log files through dozens of Hadoop jobs in order to build the data we need for our predictive models. Some tools have proven to be invaluable time savers. The tool we use most often to write map/reduce jobs is Pig, a high-level language that makes it easy to describe common map/reduce workflows. The tool builds a standalone JAR file out of a script, that eventually runs as a standard Hadoop job..
read more.........http://labs.umbrella.com/2013/04/08/pig-jruby/
read more.........http://labs.umbrella.com/2013/04/08/pig-jruby/