copy files only in dir 2 that are not also in dir 1 into a new dir

jr_bob_dobbs · 03-20-2018, 05:37 PM

Sorry for the wordy title. Could not figure out how to reduce it further without removing all meaning.

Consider a directory tree at /over/here, which contains regular files, sym links and directories, which themselves contain regular files, sym links and directories, etc.

Consider the same sort of directory tree at /over/there

My task is to create a directory /new/dir which contains everything in the second directory that is not also in the first, preserving the relative directory structure.

I think this is a simple task but I just cannot figure out how to do it with the usual tools (diff, sort, sed etc.).

Thank you.

keefaz · 03-20-2018, 05:56 PM

Maybe you could try to clone /over/there to /new/dir with rsync using exclude option
Now the job becomes to just build the exclude list (it accepts wildcards...)

BW-userx · 03-20-2018, 08:09 PM

you could write a script to copy the necessary from the first source , or command line copy first source then a script to have it go into the second source, using if statement to check first if the same file is already there or not, if it is there then do not copy that from the second source, if it is not there then copy it over into the same destination.

though in reality if it is a copy or move, I'd just copy or move the first source over then the second over top of it, it will still end result be no dups because it will just copt over top of what is already there and add the what is not already as well.

but a script would be something like this. it is not testing for or coping links. it's just a basic working script to test if the file is already there or not, and report back the results of its findings.

Code:

#!/bin/bash

working_dir="/home/userx/Documents"
destanation=/home/userx/Documents

while read f ;
do

    c=$f
    xpath=${c%/*} 
    xbase=${c##*/}
    xfext=${xbase##*.}
    xpref=${xbase%.*}
    path=${xpath}
    pref=${xpref}
    ext=${xfext}
	#covers base and sub dir
	#replaces source path with destanation then
	#adds the sub dir to match source if a match is there. 
	search=${path/$working_dir/$destanation}
     
    if [[ "$path"/"$pref" == "$search"/"$pref" ]] ; then
    {
		echo "Match"
	}
	else
	{
		echo "Not Match"
	}
	fi
done <<<"$(find "$working_dir" -type f)"

MadeInGermany · 03-21-2018, 01:39 PM

The following script does
list source files | exclude files | copy files

Code:

#!/bin/sh
excludir=/over/here
srcdir=/over/there
destdir=/new/dir

mkdir -p "$destdir" || exit
cd "$srcdir" || exit
find . -print | (
cd "$excludir" &&
while IFS= read -r f
do
  test -e "$f" || printf "%s\n" "$f"
done
) |
cpio -pd "$destdir"

syg00 · 03-21-2018, 06:10 PM

Don't see cpio getting a run much any more ...

Maybe use print0 for safety ?.

MadeInGermany · 03-22-2018, 02:21 AM

You mean special file names?
I have edited my last post, now the read handles leading/trailing space characters and backslashes in filenames.
Only newline characters in filenames are still not handled, that would denote besides a -print0 in find an adaption in read, printf, cpio.

allend · 03-22-2018, 10:35 AM

As a one-liner run from the ./over directory

Code:

mkdir new; cp ./there/$(join -v 2 <(cd here;find . ;cd ..) <(cd there;find . ;cd ..)) ./new/

Sorry - Epic fail with symlinks