At (info "(coreutils) sort invocation") it says For example, ‘sort -n -u’ inspects only the value of the initial numeric string when checking for uniqueness, whereas ‘sort -n | uniq’ inspects the entire line. *Note uniq invocation::.
OK, but you still need to add a warning about data loss. Here's a shell script: k="3 Billy 17 Villy 4 Nibblesberg 3 Philbert 3 Billy" c=sort echo We sort the students [$c] echo "$k" | $c c="sort --numeric-sort" echo Oh my gosh, we must use [$c] echo "$k" | $c c="sort --numeric-sort --unique" echo Yuck, let\'s eliminate the duplicates too [$c] echo "$k" | $c echo Oops, we caused \"data loss\". Good thing we noticed it. c="sort --unique" d="sort --numeric-sort" echo Let\'s try it the right way: [$c \| $d] echo "$k" | $c | $d Running it shows: We sort the students [sort] 17 Villy 3 Billy 3 Billy 3 Philbert 4 Nibblesberg Oh my gosh, we must use [sort --numeric-sort] 3 Billy 3 Billy 3 Philbert 4 Nibblesberg 17 Villy Yuck, let's eliminate the duplicates too [sort --numeric-sort --unique] 3 Billy 4 Nibblesberg 17 Villy Oops, we caused "data loss". Good thing we noticed it. Let's try it the right way: [sort --unique | sort --numeric-sort] 3 Billy 3 Philbert 4 Nibblesberg 17 Villy Sure, you might say, "That's already mentioned" (in the fine print). "The reader just needs to put 2 + 2 together in their heads." Yes, but anyway, the document needs to drive home the point more. Maybe the man page should say so too.
