文件操作

     sheet.txt文件内容

Tom,8
Jack,7
Tyke,3

      读取文件

def file_op():
    with open('sheet.txt', 'r') as f:
        lines = f.readlines()
        for line in lines:
            name, age = line.rstrip().split(',')
            print('{} is {} years old'.format(name, age))

      open()第一个参数是文件名，第二个参数是模式。文件的模式一般有四种，读取(r)，写入(w)，追加(a)和读写(r+)。如果希望按照二进制数据读取，则将文件模式和b一起使用（wb, r+b…）

      要读取文件内容，并把年龄和名字的顺序交换存成新文件age_name.txt，这时可以同时打开两个文件：

with open('name_age.txt', 'r') as fread, open('age_name.txt', 'w') as fwrite:
    line = fread.readline()
    while line:
        name, age = line.rstrip().split(',')
        fwrite.write('{},{}\n'.format(age, name))
        line = fread.readline()

       把对象进行序列化，那么可以考虑用pickle

def serialize():
    lines = ["I'm like a dog chasing cars.",
             "I wouldn't know what to do if I caught one...",
             "I'd just do things."]
    with open('lines.pkl', 'wb') as f:
        pickle.dump(lines, f)
    with open('lines.pkl', 'rb') as f:
        lines_back = pickle.load(f)
    print(lines_back)

异常处理

　　在深度学习中，尤其是数据准备阶段，常常遇到IO操作。这时候遇到异常的可能性很高，采用异常处理可以保证数据处理的过程不被中断，并对有异常的情况进行记录或其他动作：

for filepath in filelist:   # filelist中是文件路径的列表
    try:
        with open(filepath, 'r') as f:
            # 执行数据处理的相关工作
            ...
        
        print('{} is processed!'.format(filepath))
    except IOError:
        print('{} with IOError!'.format(filepath))
        # 异常的相应处理
        ...