copy files along with their (last) folders - linux

Question

I can find and copy all the files to a given folder using find -exec command.
But what I need to do is to find and copy all the files within a given path along with its folder in which it has been saved. So ....
/path/to/file/is/abc.txt
/another/file/is/here/xyz.txt
I need to copy these 2 files along with their path to the following folder:
/mysql/data/
The new file structure will look like this...
/mysql/data/is/abc.txt
/mysql/data/here/xyz.txt
This is done in order to avoid possible overwrite of duplicate file names. The last folder names will be unique but file names may be the same.
What is the best way to do this?

Answer 1

Here's a concise script with a rather long explanation* to accompany it.
for oldpath in $your_file_list; do
mv ${oldpath} /mysql/data${oldpath##$(dirname $(dirname $oldpath))}
done
How it works
The dirname utility removes everything up to and including the last forward slash (/) from a path. Invoking it twice will remove everything up to and including the second-to-last slash.
The idiom $(command with params) executes command with the parameters with params and returns the output.
The idiom ${var##prefix} returns the contents of the variable var with prefix removed.
Step-by-step Analysis
If oldpath is /path/to/file/is/abc.txt, then:
dirname $oldpath is /path/to/file/is
dirname $(dirname $oldpath) is /path/to/file
${oldpath##$(dirname $(dirname $oldpath))} is /is/abc.txt
which is the portion of the original path that will be appended to the new path.
* Elegant (adj.) software: any software that implements an algorithm, whose explanation is longer than the implementation itself.

Answer 2

You're going to have to script/program this solution.
Quick python example follows:
import os
import shutils
src_root = '/path/to/walk/'
dst_root = '/mysql/data/'
for root,dirs,files in os.walk(src_root):
for file in files:
dst_path = os.path.split(root)[1]
dst_path = os.path.join(dst_root, dst_path)
os.makedirs(dst_path)
src = os.path.join(root,i file)
dst = os.path.join(dst_path, file)
shutils.copyfile(srd, dst)

Answer 3

This might work for you:
a=/mysql/data
sed 's|.*\(/[^/]*/[^/]*\)|mv -v & '"$a"'\1|' file
mv -v /path/to/file/is/abc.txt /mysql/data/is/abc.txt
mv -v /another/file/is/here/xyz.txt /mysql/data/here/xyz.txt
Study the output and if all OK, then run:
sed 's|.*\(/[^/]*/[^/]*\)|mv -v & '"$a"'\1|' file | bash

Answer 4

If you want to copy a file first you need to create a directory for it. You can do that using single find command, but I'm not sure about efficiency of this solution?
#!/bin/bash
# $1 - destination
find $1 -type f -exec bash -c '
dest="$2"; # $2 is second argument passed to the script
dir=$(basename $(dirname $1));
mkdir $dest/$dir 2>/dev/null;
cp $1 "$dest/$dir/";
' -- {} $2 \; # {} = $1 in bash '..' and $2=$2
usage: ./copy copy_from copy_to
edit:
that looks better:
#!/bin/bash
dest=$2
from=$1
# copy_file from dest
copy_file() {
dir=$(basename $(dirname $from))
mkdir $dest/$dir
cp $from $dest/$dir
}
find $from -type f | while read file; do copy_file $file $dest; done