理解 os.walk Python

Question

I'm trying to walk over the directory structure and create a similar structure (but not identical).

I got confused of the use of os.path.join, the following code with 2 or more directory depth works perfectly.

DIR_1 :

A | file2.txt

B | file3.txt

file1.txt

inputpath = DIR_1
outputpath = DIR_2
for dirpath, dirnames, filenames in os.walk(inputpath):
    structure = os.path.join(outputpath, dirpath[len(inputpath):])
        for f1 in filenames:
        f = os.path.splitext(f1)[0]
        path = structure + '/' + f
        print ("The path is: ", path)
        file1 = path + '/' + f1
        print ("The file path is: ", file1)
        file_dir = dirpath + '/' + f1;
        print ("The file dir path is: ", file_dir)
        print ("\n")

But in case of just one level of depth, it add additional '/'. Is there a way to avoid this?

For example the following gives:

The path is:  DIR_2//file1
The file path is:  DIR_2//file1/file1.txt
The file dir path is:  DIR_1/file1.txt


The path is:  /A/file2
The file path is:  /A/file2/file2.txt
The file dir path is:  DIR_1/A/file2.txt


The path is:  /B/file3
The file path is:  /B/file3/file3.txt
The file dir path is:  DIR_1/B/file3.txt

Edit 1:

The output directory DIR_2 structure is similar to the original Dir_1 but not identical.

The DIR_2 should have additional one level of directory of the filename; for example rather than just

DIR_2/file1.txt

it should be

DIR_2/file1/file1.txt.

DIR_2/A/file2/file2.txt. Similarly.

Edit 2: I also need to read the content of the dirpath (of DIR_1) and select relevant text to put in the corresponding output file (of DIR_2). So i can't ignore it.

Answer 1

You should not worry about the dirpath , use it only to get the original files, all information to recreate the directory structure you already have in dirnames . The code to recreate the file structure can look like this:

for root, dirs, files in os.walk( input_path ) :
    offset = len(input_path)
    if len(root) > len(input_path) :
        offset += 1   # remove an extra leading separator

    relative_path = root[offset:]

    for d in dirs :  # create folders
        os.mkdir( os.path.join( output_path, relative_path, d )


    for f in files : # copy the files
        shutil.copy( os.path.join( root, f),
                     os.path.join( output_path, relative_path, f))

And that's it!