How to list files from n biggest directories?

I am interested in finding n biggest subdirectories (eg. 3) in my current directory AND THEN listing all the files from them. I am not interested in finding biggest files. I know there are a lot of solutions for finding biggest files and that's not what I want.

I have found my 3 biggest subdirectories with line:

$ du -hs */ | sort -rh | head -3

with result:

212K    04/
52K     02/
20K     03/

but somehow I am not able to list files from these directories (piping with ls did't work). Any suggestions how to do that?

Solution 1:

You can use the results of your piped commands with a while loop:

du -hs */ | sort -rh | head -3 | while read -r size dir 
do
   ls -l "$dir"
done

As a one liner:

du -hs */ | sort -rh | head -3 | while read -r size dir ; do ls -l "$dir"; done

Thanks to steeldriver for suggesting the use of read in a while loop over a for loop using awk, to handle filenames with spaces and special characters in a more robust manner.

Solution 2:

Command substitution, with help from awk to get first 3 directories:

ls -l $(du -hs */ | sort -rh | awk 'NR==4{exit} {print $2}')

Assuming no directory name contains any whitespace or any unusual character.

Robust approach: handling any possible file (directory) name, outputting directory name as NUL separated in awkand taking help from xargs to deal with each directory:

du -0hs */ | sort -zrh | awk 'BEGIN{RS="\0"} NR==4{exit} {printf("%s\0", $2)}' | \
    xargs -0 ls -l

Each directory content at a time:

du -0hs */ | sort -zrh | awk 'BEGIN{RS="\0"} NR==4{exit} {printf("%s\0", $2)}' | \
    xargs -0 -I{} ls -l {}