Moving A Specified Number Of Files From One Directory To Another in Linux Redhat

Question

Does anyone actually know how to move X files at a time from one directory to another? Every 'example' on this page displays an error and is useless. I can only move fifty (50) files at a time, then 'sleep' for a period of time. The file names are all different but the extension is the same. As I've stated, every 'ls' example produces an error within Redhat.

Answer 1

To move a given number of files at a time and sleeping 5 seconds in between moves you can use something similar to the following one-liner while in the source directory:

 while true; do mv -v $(ls -1 *yyy | head -n 5) otherdir/ ; sleep 5; done

Where yyy equals your common file extension, -n 5 represents moving 5 files at a time, and otherdir is your target directory. You will need to adjust the loop to suit your requirements, as listed here it will throw errors once all of the source files have been moved.

Answer 2

A small script, that could be adjusted to accept the extension and the destination directory, will accomplish this task:

#!/bin/bash

files=$(find . -maxdepth 1 -type f -name "*.txt" | head -n50)

while [[ ! -z $files ]]; do
  for f in $files; do
    mv $f ../t2
  done

  files=$(find . -maxdepth 1 -type f -name "*.txt" | head -n50)
done

Tested against a directory where I created 1,067 files, and them moved them.

I'm sure there is a way to optimize this script as well.

Edit:

As noted by a comment, the above script does not account for spaces. An approach that accounts for spaces (but not, perhaps, every special character, especially, perhaps carriage returns in the file names), might look like:

#!/bin/bash

function move50() {
  local NUM=0
  find . -name "*.txt" | head -n50 | (while read -r f; do
     echo "   $f"
      mv "$f" ../t2
      (( NUM++ ))
  done
  return $NUM)
}


cntr=1
while [[ $cntr -gt 0 ]]; do
  move50
  cntr=$?
done

This approach will handle files with spaces in the names.