bash脚本，根据目录名称的模式在目录上递归运行命令

Question

I am trying to write a small bash script to help sort some recovered directories/files back to their original file structure prior to the recovery. In a nutshell I have recovered a pool of disks that were pooled using aufs union file system, after the recovery of deleted directories/files I noticed that in a lot of the directories there are directories named ".wh..wh.original_dir_name.4_random_characters" for example ".wh..wh.Images.1E7W". This directory would be located in the same directory where the actual "Images" directory is located. These .wh directories are referred to as white space and are created by the aufs file system usually during delete operations but I noticed there are a lot of recoverable files/directories located in these but I would like to integrate them back into the normal structure.

To paraphrase, this is what I want to do but do not know how to piece it together in a bash script.

I need recursively traverse the structure, when I find a directory named something like .wh..wh. then I need to parse the real name of the directory and if this directory already exists in the same directory then execute cp -r --backup=numbered .wh..wh./* real_directory_name/ and then delete the .wh..wh directory or if the directory does not exist then simply rename the directory using the mv command. I assume that I would use the find command to traverse the structure but again not sure how to piece it altogether. Any help would be awesome.

Here is an example file structure to clarify things:

    /home/
    /home/.wh..wh.Images.67TY/
    /home/.wh..wh.Images.67TY/Camping/
    /home/Images/
    /home/Images/Camping/

In the above example I would like to use the cp command to copy the contents of /home/.wh..wh.Images.67TY to /home/Images and then delete /home/.wh..wh.Images.67TY

Answer 1

Here is a solution using only bash parameter expansion:

#!/bin/bash

for BACKUPDIR in $(find . -type d -name ".wh..wh.*")
do
    BASE=$(dirname "${BACKUPDIR}")

    LASTPART=$(basename "${BACKUPDIR}")
    D1=${LASTPART#.wh..wh.}
    D2=${D1%.*}

    if [ -d "${BASE}/${D2}" ]
    then
         cp -a "${BACKUPDIR}/"* "${BASE}/${D2}"
    else
         mv "${FILEPATH}" "${BASE}/${D2}"
    fi

done

Haven't tested the copy/move part; you should first see with echo, if this gives you, what you want.

Answer 2

#!/bin/bash
# handles directory names with spaces
IFS='
'
for BACKUPDIR in $(find . -depth -type d -name ".wh..wh.*")
do
    BASE=$(dirname "${BACKUPDIR}")
    LASTPART=$(basename "${BACKUPDIR}")
    D1=${LASTPART#.wh..wh.}
    D2=${D1%.*}

    if [ -d "${BASE}/${D2}" ]
    then
       echo copying from "${BACKUPDIR}" to "${BASE}/${D2}"
       # prevents deleting if the cp command fails
       if cp -a --backup=numbered "${BACKUPDIR}/"* "${BASE}/${D2}"
       then
           echo deleting "${BACKUPDIR}"
           rm -rf "${BACKUPDIR}"
       fi
    else
        echo moving "${BACKUPDIR}" to "${BASE}/${D2}"
        mv "${BACKUPDIR}" "${BASE}/${D2}"
    fi
done