From https://unix.stackexchange.com/a/277707/674
find . ! -empty -type f -exec md5sum {} + | sort | uniq -w32 -dD
can find duplicate files under the current directory.
What does -dD mean to uniq?
I saw the meanings of -d and -D in manpage, but not sure what they mean when they are used together. Thanks.
TLDR Bottom line, they do nothing when used together; -dD is identical to -D.
If you look at the case/switch logic of the uniq.c command you can see this first hand:
case 'd':
output_unique = false;
output_option_used = true;
break;
case 'D':
output_unique = false;
output_later_repeated = true;
if (optarg == NULL)
delimit_groups = DM_NONE;
else
delimit_groups = XARGMATCH ("--all-repeated", optarg,
delimit_method_string,
delimit_method_map);
output_option_used = true;
break;
The way that this code is structured, if either -dD is set, ouput_unique is set to false; but more importantly, output_later_repeated is set to true.
Once that condition is set, output_later_repeated, there's no feasible way for -dD to have anything but identical output to -D.
Incidentally, the computerhope man page has a better table that explains the -d and -D switches.
The uniq -dD makes no sense - it's equivalent of uniq -D. Both -dD and -D will produce always same result since -d's output set is a subset of -D.
