Variable in find command set to new variable in bash

Question

Is it possible to do something like this:

inputNum="$1"

files=($(find /dir/to/check -mtime $inputNum))

Basically the idea is that I can use an input parameter to set the number of days to find files and set it to a variable array. I am not sure on the syntax to make this readable in bash.

See Why is looping over find's output bad practice? – Stéphane Chazelas Aug 09 '17 at 14:24 — Stéphane Chazelas, Aug 09 '17 at 14:24

score 2 · Accepted Answer · answered Aug 09 '17 at 14:30

The output of find is not post-processable reliably unless you use -print0 instead of -print (-print implied when no action is specified).

To post-process the output of find -print0 and store the file paths in an array:

With bash4.4+:

readarray -td '' files < <(find /dir/to/check -mtime "$inputNum" -print0)

With older versions:

files=()
while IFS= read -rd '' file; do
  files=("${files[@]}" "$file")
done < <(find /dir/to/check -mtime "$inputNum" -print0)

More generally, you'd want to read the recommendations at: Why is looping over find's output bad practice?

score 0 · Answer 2 · answered Aug 09 '17 at 22:48

If you can use zsh instead of bash, then it's a lot easier. Zsh can match files based on their attributes through glob qualifiers.

files=(/dir/to/check/**/*(Dm$inputNum))

**/* searches subdirectories recursively. The parentheses after * contain glob qualifiers: D to include dot files, and m to match files based on their modification time in days like find -mtime (you can also choose a different unit and make the time a limit, e.g. (Dmm-$min) acts like find -mmin -$min).

Zsh is preinstalled on macOS. On most other unix systems (in particular most Linux distributions), it's available as a package but not installed by default.

Variable in find command set to new variable in bash

2 Answers2