killerwhile · February 1, 2016 20:33 · killerwhile · Feb 1, 2016
diff --git a/gistfile1.sh b/gistfile1.sh
 # hadoop classpath prior to 2.6 returns wildcarded path (starting hadoop 2.6, hadoop classpath --glob is doing the trick by itself)
 # It's sometimes convenient to have a nicely printed list of jars that that part of the classpath when running hadoop jar.

 for i in $(hadoop classpath | sed -e "s/:/ /g"); 
 do 
    echo $i | egrep "\.jar$" >/dev/null && python -c "import os,sys; print os.path.realpath(sys.argv[1])" $i
 done | sort | uniq


 # If you want to generate the list of ArtifactIds to exclude from your packaging, use this version

 for i in $(hadoop classpath | sed -e "s/:/ /g");
 do
    echo $i | egrep "\.jar$" >/dev/null && basename $(python -c "import os,sys; print os.path.realpath(sys.argv[1])" $i) | sed 's/-[0-9].*//' | grep -v "\.jar";
 done | sort | uniq | paste -sd,
	# hadoop classpath prior to 2.6 returns wildcarded path (starting hadoop 2.6, hadoop classpath --glob is doing the trick by itself)
	# It's sometimes convenient to have a nicely printed list of jars that that part of the classpath when running hadoop jar.

	for i in $(hadoop classpath \| sed -e "s/:/ /g");
	do
	echo $i \| egrep "\.jar$" >/dev/null && python -c "import os,sys; print os.path.realpath(sys.argv[1])" $i
	done \| sort \| uniq


	# If you want to generate the list of ArtifactIds to exclude from your packaging, use this version

	for i in $(hadoop classpath \| sed -e "s/:/ /g");
	do
	echo $i \| egrep "\.jar$" >/dev/null && basename $(python -c "import os,sys; print os.path.realpath(sys.argv[1])" $i) \| sed 's/-[0-9].*//' \| grep -v "\.jar";
	done \| sort \| uniq \| paste -sd,