Question

I have a set of files, each containing a single (integer) number, which is the number of files in the directory of the same name (without the .txt suffix) – the result of a wc on each of the directories.

I would like to sum the numbers in the files. I’ve tried:

i=0; 
find -mindepth 1 -maxdepth 1 -type d -printf '%f\n' | while read j; do i=$i+`cat $j.txt`; done
echo $i

But the answer is 0. If I simply echo the output of cat:

i=0; find -mindepth 1 -maxdepth 1 -type d -printf '%f\n' | while read j; do echo `cat $j.txt`; done

The values are there:

1313
1528
13465
22258
7262
6162
...

Presumably I have to cast the output of cat somehow?

[EDIT]

I did find my own solution in the end:

i=0; 
for j in `find -mindepth 1 -maxdepth 1 -type d -printf '%f\n'`; do 
    expr $((i+=$(cat $j.txt))); 
done; 

28000
30250
...
...
647185
649607

but the accepted answer is neater as it doesn’t output along the way

Answer 1

The way you’re summing the output of cat should work. However, you’re getting 0 because your while loop is running in a subshell and so the variable that stores the sum goes out of scope once the loop ends. For details, see BashFAQ/024.

Here’s one way to solve it, using process substitution (instead of pipes):

SUM=0
while read V; do
    let SUM="SUM+V" 
done < <(find -mindepth 1 -maxdepth 1 -type d -exec cat "{}.txt" \;)

Note that I’ve taken the liberty of changing the find/cat/sum operations, but your approach should work fine as well.