Copy first 100 regex matched filenames to another directory in unix

Question

My use case is to copy all filenames matching regex [0-9]{10} to a new directory. I also have to restrict the copied files to 100. I tried going through a few sources that explains how we can use regular expressions to do this but my limited understanding of bash and unix is limiting me from getting this usecase working. I tried something similar to this: How to copy multiple files from a different directory using cp?

Any help will be highly appreciated.

Answer 1

bash: store all the files in an array, then take a slice of the first 100 elements

all_files=( [0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9] )  # glob pattern, not regex
cp -t /destination/dir "${all_files[@]:0:100}"

Answer 2

Something like this should work for you:

cp `ls -1 | egrep '[0-9]{10}' | head -100` <destination directory>

(Depending on your system, you might have a different grep command, or one that requires using the -e switch)

Answer 3

While I think Glenn Jackman's answer is one of the nicest I've seen in a while, if you really do need to use a regex, then pathname expansion in an array won't work for you. Instead, you can either use find to find files (and populate your array), or you can step through a directory and use the regex matching built in to bash.

---

First strategy, using find (per Greg's BashFAQ/020):

unset files i
while IFS= read -r -d $'\0' file; do
  files[i++]="$file"
done < <(find -E ./ -type f -regex '\./[0-9]{10}' -print0)

Note that find's -regex has implicit ^ and $ anchors. I'm using -E to tell find that I want to use ERE instead of BRE (which works in macOS, FreeBSD, other BSDs...). In Linux, you may want to use the -regextype option ... or just express yourself in BRE.

Then select just the first 100 array items as Glenn suggested:

cp "${files[@]:0:100}" /path/to/destination/

---

The second strategy, using Bash's built-in regex support, might be done with a bit of scripting:

unset n
for file in *; do
  [[ $file =~ ^[0-9]{10}$ ]] &&
  mv -v "$file" dest/ &&
  (( ++n >= 100 )) && break
done

This uses globbing to identify all files, and then for each one that matches your regex, it moves the file and increments a counter. The increment also checks to see if it's exceeded your threshold and if so, breaks out of the loop.

You could make it a one-liner if you like. I did when writing and testing it. And this could of course be written longer if you don't like your scripts terse. :)

Answer 4

with this command you can find all the files you need : $sudo find / -regextype sed -regex "[0-9]{10}"

*note that you can replace / with any directory that you want to search in for your needle

and with this you are going to do what you want :

$sudo find / -regextype sed -regex "[0-9]{10}" -exec cp {} /path/you/want \;

Answer 5

Thanks for the valuable suggestions. I came up with following solution :

ls -1 | egrep [0-9a-f]{10} | head -100 | xargs -I{} cp -f {} <your directory>