How to modify a find function loop to handle backslashes and whitespaces?

Question

Currently, I am trying to parse out all the files inside a directory. I have a find function, but it appears to not be able to parse in file directories with whitespaces into a loop. Here, "$DIR" is the directory I wish to search in. Does anyone have any ideas how I can modify it to work? thanks.

for file in $(find "$DIR" -type f)
do
   echo "$file"
done

Related: https://unix.stackexchange.com/questions/321697/why-is-looping-over-finds-output-bad-practice — Kusalananda, Jun 08 '18 at 18:19

score 5 · Answer 1 · answered Jun 08 '18 at 17:41

5

You don't need to for loop this.

find "$dir" -type f

will by default output all found objects.

You can make this explicit with:

find "$dir" -type f -print

If you really want to iterate over these, use a null separator and xargs:

find "$dir" -type f -print0 | xargs -0 -n1 echo

Or find's own -exec:

find "$dir" -type f -exec echo "{}" \;

answered Jun 08 '18 at 17:41

DopeGhoti

76,081

If I wanted to pass the filenames found into a function, is there a standard way to do it? – user321627 Jun 08 '18 at 17:49
I'd read the results of find into an array (arr=( $( find [...] ) )), and reference that array within the function, or pass its contents to the function func "${arr[@]}". – DopeGhoti Jun 08 '18 at 17:52
1

@DopeGhoti isn't that going to have the same word-splitting issues as the original? See for example How to pass files found by find as arguments? – steeldriver Jun 08 '18 at 19:18
Yeah, probably. Let me revise. I'd probably pass the directory to the function, and let it handle the finding. – DopeGhoti Jun 08 '18 at 21:49

score 5 · Accepted Answer · answered Jun 08 '18 at 18:13

find ... | xargs ... and find ... -exec ... are both better options than this: to use a shell loop to iterate over find results correctly, we must use a while read loop:

while IFS= read -d '' -r filename; do 
    echo "$filename"
done < <(find "$dir" -type f -print0)

A lot to unpack there:

we use a Process Substitution to execute the find command and be able to read from the results like it's a file.
To read a line of input verbatim, the bash idiom is IFS= read -r line. That allows arbitrary whitespace and backslashes to be read into the variable.
read -d '' uses the null byte (as produced by -print0) as the end-of-line character instead of newline.

Also, don't use ALL_CAPS_VARNAMES: it's too easy to make mistakes — glenn jackman, Jun 08 '18 at 18:16

score 4 · Answer 3 · edited Jun 08 '18 at 18:18

4

You should not parse the output of find for the same reasons you shouldn't parse ls

Try using find/exec instead:

find "$DIR" -type f -exec echo {} \;

edited Jun 08 '18 at 18:18

Kusalananda

333,661

answered Jun 08 '18 at 17:42

jesse_b

37,005

How to modify a find function loop to handle backslashes and whitespaces?

3 Answers3