How to find a file and copy its directory without copying the other files in that same directory?

Question

I have a lot of different folders with pdf files and other extensions. I want to copy all the pdf files to another destination. But if I do that with find like this:

"${SEARCHDIR}" -iname "*.pdf" | xargs -d "\n" -I {} cp -r --backup=t {} "${OUTPUTFOLDER}/"

I get all the pdf files in single destination and that is too messy for me. For that reason I want to copy the matching files including the parent directory but WITHOUT the non-matching files in that directory.

Input:

user/search/test/sub
user/search/test/sub2
user/search/test/file.pdf
user/search/test/file.doc
user/search/test2/sub
user/search/test2/sub2/file2.pdf
user/search/test2/file3.pdf
user/search/test2/file.doc

Should be copied as

/destination/test/file.pdf
/destination/sub2/file2.pdf
/destination/test2/file3.pdf

EDIT: I don't want to copy the whole path. I don't want my output to be like this:

/destination/test/file.pdf
/destination/test2/sub2/file2.pdf
/destination/test2/file3.pdf

How to do this?

Answer 1

In zsh, you could do:

set -o extendedglob
for file ( $SEARCHDIR/**/*.(#i)pdf(ND^/) ) {
  dest=$OUTPUTFOLDER/${file:t2}
  mkdir -p -- $dest:h && cp --backup=t -- $file $dest
}

Where ${file:t2} gets you the two-component tail of the file, that is the file name and that of its parent directory. Then we create the head of that path as a directory and copy the file there.