#!/bin/sh # every directory cas/btih/{btih}/ # should contain exactly one subdirectory # or exactly one file zcat cas.names.txt.gz | awk -F/ ' { parent = $1"/"$2"/"$3 count[parent]++ lines[parent] = lines[parent] $0 "\n" } END { for (p in count) if (count[p] > 1) printf "%s", lines[p] } '