Add ability to name mapreduce job generated by Sqoop

Names of generated mapreduce job is currently the same as jar name (for example table_name.jar). It would be nice if Sqoop would have command line parameter to override that so that user could modify that arbitrarily.

Please note that this can be partially achieved even now, by passing Dmapred.job.name. However this approach do not work reliably when executing from Oozie or other tools that are directly executing hadoop jar.