Getting the total size of my five largest files in bash

I am trying to get the total size of my five largest files found in a directory but I am not getting du to work on my list. I have two ways of finding and sorting, the five largest files.

#!/bin/bash

DIR=$1 #The starting directory path


if [ ! -d "$DIR" ]; then #if the directory is not found
    echo "The directory doesn't exist!" 
    exit 1
fi

echo "Five largest files using ls are:"
test_ls=$( ls -lhR "$DIR" | grep '^-' | sort -r -k 5 -h | head -n 5 )
du -ch "$test_ls"


echo "Five largest files using find/DU are:"
test_find=$( find "$DIR" -type f -exec du -ch {} + | sort -rh | head -n 5 )
du -ch "$test_find"

echo "Total number of files: "
ls -lhR "$DIR" | grep '^-' | wc -l
echo "Total size of files: "
du -sh "$DIR" | awk '{print $1}'

If I apply du on the ls variable I get :

du: invalid option -- 'r'
du: invalid option -- 'w'
du: invalid option -- '-'
du: invalid option -- 'r'

and if I apply it on the find variable I get this for each of the five files

du: cannot access '429M': No such file or directory

Both the ls version and find version works fine for listing the five largest files under the given directory, but I am relly lost in how I could add their sizes together.

Not sure how best to fix your code, I would propose a different approach without the need for du as find can return file sizes with -printf '%s'.

Note also, that you should never parse ls and that find unlike ls will run recursively if you don't limit that with -maxdepth option.

This would find the five largest files in a directory and sum its sizes:

find "$DIR" -maxdepth 1 -type f -printf '%s\n' | sort -nr | head -n5 | paste -sd+ | bc | numfmt --to=iec

• find ... -printf '%s\n' will print file sizes in bytes
• | sort -nr | head -n5 will find the five largest numbers
• | paste -sd+ will join the numbers with a plus sign in between them, so it's a mathematic expression
• | bc will run the mathematic expression
• | numfmt --to=iec (optional) will convert the size from bytes to human-readable format.

To get additional information, you can save find output e.g. in an array:

DIR=/some/dir
file_sizes=($(find "$DIR" -maxdepth 1 -type f -printf '%s\n' | sort -nr))

num_files=${#file_sizes[@]}
total_size=$(IFS=+; echo "$((${file_sizes[*]}))" | numfmt --to=iec)
biggest_files=$(IFS=+; echo "$((${file_sizes[*]:0:5}))" | numfmt --to=iec)

printf 'Total number of files: %d\nTotal size of files: %s\nSize of biggest 5: %s\n' $num_files $total_size $biggest_files