Python-批量重命名

数据集来源：http://download.tensorflow.org/example_images/flower_photos.tgz

命名

命名前

代码

import os

train_image_path=r'D:\Projects\DeepLearning\Dataset\flower_photos\train'
flower_class=[flower for flower in os.listdir(train_image_path)]

num=0
for flower in flower_class:
    flower_path=train_image_path+'\\'+flower
    for file in os.listdir(flower_path):
        os.rename(os.path.join(flower_path,file),os.path.join(flower_path, str(num)+'.jpg'))
        num+=1

命名后

参考1

参考2

遇到问题

改名后发现继承Dataset类，重写__len__ 和__getitem__时，使用__getitem__时，发现遍历顺序有问题。
如果用循环给文件命名，则文件名就会是1,2,3…，10,11,12,13…，100,101…，但是遍历这些文件时，顺序就会变成1,10,100,101，…109，11，…，19，…，2,20,200，…，那么如何按1,2,3,4…的顺序遍历呢？
解决方法：
可以把原来的文件重命名，在原文件名前面加上0，例如00001,00002，…，这样就可以按顺序遍历了。
注： 这里我先将图片名命名为0,1,2…，没有 前面的花名，以便于操作。（实际上是菜）

代码

import os

train_image_path=r'D:\Projects\DeepLearning\Dataset\flower_photos\train'
flower_class=[flower for flower in os.listdir(train_image_path)]

for flower in flower_class:
    flower_path=train_image_path+'\\'+flower
    for file in os.listdir(flower_path):
        name = file.split('.')[0]
        os.rename(os.path.join(flower_path,file),os.path.join(flower_path, '%05d' % int(name)+'.jpg')) 

#‘%05d’表示一共5位数

现在遍历就是按照顺序了，2就是00002.jpg，不再是10.jpg了。

参考

整合

import os

def ReNameFlower(image_path,flower_class):
    num=0
    for flower in flower_class:
        flower_path=image_path+'\\'+flower
        for file in os.listdir(flower_path):
            os.rename(os.path.join(flower_path,file),os.path.join(flower_path, str(num)+'.jpg'))
            num+=1
    for flower in flower_class:
        flower_path=image_path+'\\'+flower
        for file in os.listdir(flower_path):
            name = file.split('.')[0]
            os.rename(os.path.join(flower_path,file),os.path.join(flower_path, '%05d'%int(name)+'.jpg'))

def main():
    train_image_path = r'D:\Projects\DeepLearning\Dataset\flower_photos\train'
    train_flower_class = [flower for flower in os.listdir(train_image_path)]
    ReNameFlower(train_image_path,train_flower_class)

    val_image_path = r'D:\Projects\DeepLearning\Dataset\flower_photos\val'
    val_flower_class = [flower for flower in os.listdir(val_image_path)]
    ReNameFlower(val_image_path,val_flower_class)

if __name__=='__main__':
    main()