How to merge directories differing only by character case in their names?

Question

Some cloud storage services let one download chosen directories as ZIP archives. But it appears that, on one of such services, the archives contain different-cased versions of the same directory. E.g. if I have Books directory stored in the cloud, the archive could have two entries for it: Books and books, and both of these may have some data – several files in one and others in another, so that total number of files is the same as in the cloud. And this problem happens recursively in the whole directory tree.

All cases I've seen only split the directories in no more than two versions: 1) correctly-named, and 2) all-lowercase.

I'd like, after (or in the process of) extracting this ZIP archive, to merge the directories, so that for each pair of case-differing directories I got only the correctly-named directory with all its original data as well as the data from its all-lowercase sister.

What is an easy way to do this?

Answer 1

To merge all directories to lowercase, one approach is to recurse over all directories and merge them into their lowercase counterpart. The traversal needs to be done in depth first mode, so that when you merge two directories, their content is already normalized.

The following script might do the job:

find . -depth -name '*[[:upper:]]*' -type d -execdir sh -c '
  source=$0
  target=$(echo "$source" | tr "[:upper:]" "[:lower:]")
  if [ "$source" != "$target" ]; then  # need to rename or merge
    if [ -d "$target" ]; then
      # merge $source (mixed- or uppercase) into $target (lowercase)
      find "$source" -mindepth 1 -maxdepth 1 -exec mv -bt "$target" {} +
      rmdir "$source"
    else
      mv "$source" "$target"
    fi
  fi
' {} \;