Time for action – building a JAR file

Before we run our job in Hadoop, we must collect the required class files into a single JAR file that we will submit to the system.

Create a JAR file from the generated class files.

$ jar cvf wc1.jar WordCount1*class

What just happened?

We must always package our class files into a JAR file before submitting to Hadoop, be it local or on Elastic MapReduce.

Tip

Be careful with the JAR command and file paths. If you include in a JAR file class the files from a subdirectory, the class may not be stored with the path you expect. This is especially common when using a catch-all classes directory where all source data gets compiled. It may be useful to write a script to change into the directory, convert the required files into JAR files, and move the JAR files to the required location.