to sort the output of hadoop fs -ls today needs you to sort it externally, which is good enough, but not very convenient because it relies on the output listing formats.
Implement hadoop fs -ls -r -t -S as a client command to help locate files which are the biggest or the most recently modified.