python - 如何在 Python 中移动目录？

Question

我需要将目录从一个位置移动到同一文件系统上的另一个位置。我知道类似的解决方案shutil.move()，但是有问题的文件系统是 SD 卡（因此非常慢），并且有很多文件要移动，所以简单地复制它们然后删除原件是不可接受的。Unixmv命令可以将目录从一个文件系统移动到同一个文件系统，而无需复制任何文件——有没有办法在 Python 中做到这一点？

Answer 1

事实证明，答案是肯定的。您可能知道，您可以使用os.rename(r'D:\path1\myfile.txt', r'D:\path2\myfile.txt'). 您可以对目录执行相同的操作：

os.rename(r'D:\long\path\to\mydir', r'D:\mydir')

但是，当然，只有在 D:\mydir 尚不存在时才有效。如果它确实存在，并且您想将已经存在的文件与您正在移动的文件合并，那么您需要变得更聪明一些。这是一个片段，可以做你想做的事：

def movedir(src, dst):
    try:
        os.rename(src, dst)
        return
    except FileExistsError:
        pass
    for root, dirs, files in os.walk(src):
        dest_root = os.path.join(dst, os.path.relpath(root, src))
        done = []
        for dir_ in dirs:
            try:
                os.rename(os.path.join(root, dir_), os.path.join(dest_root, dir_))
                done.append(dir_)
            except FileExistsError:
                pass
        for dir_ in done:
            dirs.remove(dir_)
        for file in files:
            os.replace(os.path.join(root, file), os.path.join(dest_root, file))
    for root, dirs, files in os.walk(src, topdown=False):
        os.rmdir(root)

这是一个带有注释的版本，解释了一切的作用：

def movedir(src, dst):
    # if a directory of the same name does not exist in the destination, we can simply rename the directory
    # to a different path, and it will be moved -- it will disappear from the source path and appear in the destination
    # path instantaneously, without any files being copied.
    try:
        os.rename(src, dst)
        return
    except FileExistsError:
        # if a directory of the same name already exists, we must merge them.  This is what the algorithm below does.
        pass
    for root, dirs, files in os.walk(src):
        dest_root = os.path.join(dst, os.path.relpath(root, src))
        done = []
        for dir_ in dirs:
            try:
                os.rename(os.path.join(root, dir_), os.path.join(dest_root, dir_))
                done.append(dir_)
            except FileExistsError:
                pass
        # tell os.walk() not to recurse into subdirectories we've already moved.  see the documentation on os.walk()
        # for why this works: https://docs.python.org/3/library/os.html#os.walk
        # lists can't be modified during iteration, so we have to put all the items we want to remove from the list
        # into a second list, and then remove them after the loop.
        for dir_ in done:
            dirs.remove(dir_)
        # move files.  os.replace() is a bit like os.rename() but if there's an existing file in the destination with
        # the same name, it will be deleted and replaced with the source file without prompting the user.  It doesn't
        # work on directories, so we only use it for files.
        # You may want to change this to os.rename() and surround it with a try/except FileExistsError if you
        # want to prompt the user to overwrite files.
        for file in files:
            os.replace(os.path.join(root, file), os.path.join(dest_root, file))
    # clean up after ourselves.
    # Directories we were able to successfully move just by renaming them (directories that didn't exist in the
    # destination already) have already disappeared from the source.  Directories we had to merge are still there in
    # the source, but their contents were moved.  os.rmdir() will fail unless the directory is already empty.
    for root, dirs, files in os.walk(src, topdown=False):
        os.rmdir(root)

movedir(r'D:\long\path\to\mydir', r'D:\mydir')

请注意，以这种方式使用 os.rename() 仅在源路径和目标路径位于同一文件系统上时才有效（在 Windows 上，如果它们具有相同的驱动器号，则为真）。如果它们位于不同的驱动器号上（即一个在 C: 上，另一个在 D: 上）或者其中一个路径包含重解析点（如果您不知道那是什么，请不要担心，您可能永远不会遇到一个），您将需要使用shutil.move()，它复制文件然后从源中删除它们 - 这是 Windows 在驱动器之间移动文件时所做的事情，并且完成所需的时间大约一样长。