python os.walk os.path glob.glob 遍历目录/文件夹 下所有文件/目录 的若干方法

比如我电脑上的一个文件下的内容如下，两个子文件中只有文件无文件夹：

os.walk遍历所有子目录和文件

首先需要理解os.walk返回的三个值的含义

root, dirnames, filenames = os.walk(filepath)

root 所指的是当前正在遍历的这个文件夹的本身的地址
dirs 是一个 list ，内容是该文件夹中所有的目录的名字(不包括子目录)
files 同样是 list , 内容是该文件夹中所有的文件(不包括子目录)

是不是感觉有点抽象，实际跑一下程序帮助大家理解

for root, dirnames, filenames in os.walk(filepath):
    print(root)

D:\9_Code\Tensorflow_tutorials
D:\9_Code\Tensorflow_tutorials.ipynb_checkpoints
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization

三个目录的绝对路径

for root, dirnames, filenames in os.walk(filepath):
    print(dirnames)

[’.ipynb_checkpoints’, ‘Convnet_Visualization’]
[]
[]

三个目录中的子目录

for root, dirnames, filenames in os.walk(filepath):
    print(filenames)

[‘cache.tf-data.index’, ‘Cat.jpg’, ‘Convnet_Visualization_CAM - 副本.ipynb’, ‘Convnet_Visualization_CAM.ipynb’, ‘Egyptian_cat.jpg’, ‘Load images.ipynb’]
[‘Load images-checkpoint.ipynb’]
[‘cat_egyptian.jpg’, ‘Convnet_Visualization.ipynb’, ‘Convnet_Visualization.png’, ‘Convnet_Visualization1.png’, ‘Convnet_Visualization2.png’, ‘Convnet_Visualization3.png’, ‘Convnet_Visualization4.png’, ‘Egptian_cat.jpg’, ‘sedan.jpg’, ‘tabby.jpg’, ‘Tabby1.jpg’, ‘tiger_cat.jpg’, ‘tigger_cat.jpg’]

三个目录中的文件

遍历所有子目录

dirlist=[]
for root, dirnames, filenames in os.walk(filepath):
    for dirname in dirnames:
    	dirlist.append(os.path.join(root, dirname))
        print(os.path.join(root, dirname))

D:\9_Code\Tensorflow_tutorials.ipynb_checkpoints
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization

遍历所有子文件

filelist=[]
for root, dirnames, filenames in os.walk(filepath):
    for filename in filenames:
    	filelist.append(os.path.join(root,filename))
        print(os.path.join(root,filename))

D:\9_Code\Tensorflow_tutorials\cache.tf-data.index
D:\9_Code\Tensorflow_tutorials\Cat.jpg
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization_CAM - 副本.ipynb
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization_CAM.ipynb
D:\9_Code\Tensorflow_tutorials\Egyptian_cat.jpg
D:\9_Code\Tensorflow_tutorials\Load images.ipynb
D:\9_Code\Tensorflow_tutorials.ipynb_checkpoints\Load images-checkpoint.ipynb
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\cat_egyptian.jpg
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\Convnet_Visualization.ipynb
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\Convnet_Visualization.png
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\Convnet_Visualization1.png
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\Convnet_Visualization2.png
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\Convnet_Visualization3.png
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\Convnet_Visualization4.png
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\Egptian_cat.jpg
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\sedan.jpg
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\tabby.jpg
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\Tabby1.jpg
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\tiger_cat.jpg
D:\9_Code\Tensorflow_tutorials\Convnet_Visualization\tigger_cat.jpg

os.path

同样我们可以利用os.path自己写循环实现相同的目的

遍历所有子目录

import os

filepath=r'D:\9_Code\Tensorflow_tutorials'
dirlist=[]
list1= os.listdir(filepath)
for file1 in list1:
    path = os.path.join(filepath, file1)
    if os.path.isdir(path):
        dirlist.append(path)
print(dirlist)

遍历所有子文件

import os

filepath=r'D:\9_Code\Tensorflow_tutorials'
filelist=[]
list1= os.listdir(filepath)
for file1 in list1:
    path = os.path.join(filepath, file1)
    if os.path.isfile(path):
        filelist.append(path)
    else:
        list2=os.listdir(path)
        for file2 in list2:
            path2 = os.path.join(path, file2)
            if os.path.isfile(path2):
                filelist.append(path2)
print(filelist)

glob.glob明确的文件类型

如果我们的检索目标比较明确，比如我要得到此目录下的.jpg文件，可以使用glob.glob

import glob

filepath=r'D:\9_Code\Tensorflow_tutorials'
jpg_img = glob.glob(filepath+'\*.jpg')
print(jpg_img)

[‘D:\9_Code\Tensorflow_tutorials\Cat.jpg’, ‘D:\9_Code\Tensorflow_tutorials\Egyptian_cat.jpg’]

便可直接返回目录下的列表形式，
注意：输入是相对路径，则返回也是相对路径

glob.glob多种文件类型

一个小技巧’*.[pn][jp]g’可以实现jpg或者png文件的同时读取

import glob

filepath=r'D:\9_Code\Tensorflow_tutorials'
jpg_img = glob.glob(filepath+'\*.[pn][jp]g')

比如写一个便利两级子目录的文件和图片的代码

import os
import glob
import csv

filepath = './'
filelist1 = sorted(os.listdir(filepath))

sum_frames=0
sum_file=0
with open('num_frame.csv', mode='w', newline='') as f:
    writer = csv.writer(f)

    for file1 in filelist1:

        if len(file1)<4:

            filelist2 = sorted(os.listdir(file1))
            for file2 in filelist2:

                path = os.path.join(filepath, file1, file2)
                frames = glob.glob(path+'/*.[jp][pn]g')
                num_frame =len(frames)
                sum_frames +=num_frame
                writer.writerow([num_frame])
                sum_file +=1
                # print(path, num_frame)
    print('totoal number frames:', sum_frames)
    print('totol file:', sum_file)

判断是否未图像

images_names = [f for f in os.listdir(videos[i] + frames_path) if
              f.endswith(('.jpg', '.jpeg', '.png'))]
images_names.sort()

检查文件夹是否存在

另外在一些创建文件夹操作中，使用os.path.exists可以避免重复操作

if not os.path.exists(filename): 
    os.makedirs(filepath)

检查是否为文件夹

for ....:
    if not example_folder.is_dir():
        continue

获得当前绝对路径

from pathlib import Path
default_data_dir = Path(__file__).resolve().parent.parent / "data"

删除路径下所有文件和文件夹

在这里插入代码片