Sync folders using diff + awk

#diff #awk #linux #admin

Real life diff use:

I want to copy only the missing 'jpg' files at the ~/tmp/wallpapers

diff -u ~/img/backgrounds ~/tmp/wallpapers

Only in /home/sergio/tmp/wallpapers: .git
Only in  /home/sergio/tmp/wallpapers: .gitignore
Only in  /home/sergio/tmp/wallpapers: README.md
Only in  /home/sergio/tmp/wallpapers: thumbs
Only in  /home/sergio/img/backgrounds: wallpaper-1791868.jpg
Only in  /home/sergio/img/backgrounds: wallpaper-2099153.jpg
Only in  em /home/sergio/img/backgrounds: wallpaper-2226037.jpg

Using awk with two field separators [: ] I only need the third and fifth fields to compose a copy command:

diff -u ~/img/backgrounds ~/tmp/wallpapers | awk -F'[: ]' '/backgrounds/ {print $3"/"$5}'

/home/sergio/img/backgrounds/wallpaper-1791868.jpg
/home/sergio/img/backgrounds/wallpaper-2099153.jpg
/home/sergio/img/backgrounds/wallpaper-2226037.jpg

Now I only have to add some strings to my awk command and pipe the result to the shell

"cp " ............... copy comand with space
$3  ................ the third fild - the basename
"/" ................ add the slash so we do not mess things up
$5 ................. the missing file name
" ~/tmp/wallpapers"  destination folder
| sh ................ pipe the preceding command to the shell

diff -u ~/img/backgrounds ~/tmp/wallpapers \
| awk -F'[: ]' '/backgrounds/ {print "cp " $3"/"$5" ~/tmp/wallpapers/"}' | sh

An interesting thing is that the backgrounds folder has no .git files on it but the wallpapers do, that's why I am filtering '/backgrounds/' with awk and not using recursion '-r', because the wallpapers folder has a thumbs subfolder