Executing script from Hadoop

[Madhumitha Baskaran]

I have a requirement in hadoop and I am not able to get it right. I have to execute a UNIX script inside hadoop program and whatever is being echoed by UNIX scripts should be obtained in the hadoop program as output from Unix script. How will it be possible?

In ordinary JAVA program, I am able to execute JAVA program and get the output of unix script. But same is not working in hadoop. Can you please help?

[Alex Holmes]

Take a look at Hadoop Streaming: http://wiki.apache.org/hadoop/HadoopStreaming

It will allow you to use UNIX scripts in combination with MapReduce. Input records are supplied to stdin of your scripts, and output records are collected from stdout.