有选择地将文件从一个文件夹目录复制到另一个文件夹

Question

I have a directory tree where names of folder matter a lot. Also i have a csv saying from folder1>folder2>folder3>foo.txt. folder1,folder2,folder3 and txt all in different column of csv. I need to keep the directory structure as is and copy the files that are given in the csv.

Approch i am trying is copied the directory tree and writing a python code to remove unwanted files. So there are lot of loops but I have over 415,000 rows in csv.

csv example:<br/>
pdf_no .   folder1. folder2 . folder3. <br/> 1 .  . abc. pqr. xyz.<br/>

This is the format of csv and I have no issue with extracting column data with help of pandas dataframe in python. Originally it was a .dta file I converted to .csv with pandas. So 'folder1' > 'folder 2' > 'folder 3' > 'pdf_no'. .csv with pandas. So 'folder1' > 'folder 2' > 'folder 3' > 'pdf_no'. The 'pdf_no.' column contains filenames which is a number that we want in the given folder maintaining file structure.

So it takes a lot of time and whenever I change a bit it again will take much time and I don't even know if it getting correct.

Answer 1

You're needing the shutil.copytree method. Here is what you could do:

1. Read your CSV
2. Build the file path (with os.path.join())
3. Use shutil.copytree to copy the file and its parent directories to the target

Maybe you will have to add a try...except block to avoid an OsError when the target file already exists, or delete the target file before copying the new one.

Answer 2

Sample csv

pdf_no,folder1,folder2,folder3
1,abc,def,ghi
2,xyz,pqr,
3,abc,def,ghi

Sample code

import csv
import os
import shutil


target_csv = 'selection.csv'
target_dir = 'selected_20190828/'
source_dir = 'original_directory/'

with open(target_csv) as f:
    rows = csv.reader(f)
    for line_no, row in enumerate(rows):
        if line_no == 0:  # Skip the first line because it's the title
            continue

        pdf_name = row[0] + '.pdf'
        dir_path = os.path.join(*row[1:])

        source = os.path.join(source_dir, dir_path, pdf_name)
        if not os.path.isfile(source):
            print('not exist: ', line_no, source)
            continue
        target = os.path.join(target_dir, dir_path)
        os.makedirs(target)
        shutil.copy2(source, target)

Explanation

You don't need pandas actually, all you need is

• csv.reader to read csv file into list
• os.makedirs to create folders (this method is similar to mkdir -p in bash)
• os.path.join
• shutil.copy2 to copy file to a new folder
• os.path.isfile to make sure the original file exists

I have tested the code above. It should be working.