python 文件流

open | write | read

>>> with open(‘demo.txt‘ , ‘wt‘) as f:   #以文本(wt)形式写入操作

...     f.write("hello guohai")

12

>>> with open(‘demo.txt‘ , ‘rt‘) as f: #以文本(rt)形式读取操作

...     data = f.read()

>>> data

‘hello guohai‘

>>> with open(‘demo.txt‘ , ‘at‘) as f: #以文本(at)形式追加操作

...     f.write(‘hello zeopean‘)

13

>>> with open(‘demo.txt‘ , ‘rt‘) as f:

...     data = f.read()

>>> data

‘hello guohaihello zeopean‘

-- print 输出到文件中

with open(‘d:/work/test.txt‘, ‘wt‘) as f:

print(‘Hello World!‘, file=f)

-- 二进制读写

如果你想从二进制模式的文件中读取或写入文本数据，必须确保要进行解码和编码

操作。比如：

 1 with open(‘somefile.bin‘, ‘rb‘) as f:
 2
 3 data = f.read(16)
 4
 5 text = data.decode(‘utf-8‘)
 6
 7 with open(‘somefile.bin‘, ‘wb‘) as f:
 8
 9 text = ‘Hello World‘
10
11 f.write(text.encode(‘utf-8‘))
12
13  

-- 判断文件是否存在

>>> import os

>>> if not os.path.exists(‘somefile‘):

... with open(‘somefile‘, ‘wt‘) as f:

... f.write(‘Hello\n‘)

... else:

... print(‘File already exists!‘)

...

File already exists!

StringIO | BytesIO

 1 -- StringIO
 2
 3 >>> s = io.StringIO()
 4
 5 >>> s.write(‘Hello World\n‘)
 6
 7 12
 8
 9 >>> print(‘This is a test‘, file=s)
10
11 15
12
13 >>> s.getvalue()
14
15 ‘Hello World\nThis is a test\n‘
16
17 >>>
18
19 >>> s = io.StringIO(‘Hello\nWorld\n‘)
20
21 >>> s.read(4)
22
23 ‘Hell‘
24
25 >>> s.read()
26
27 ‘o\nWorld\n‘
28
29 >>>
30
31
32
33 -- BytesIO
34
35 >>> s = io.BytesIO()
36
37 >>> s.write(b‘binary data‘)
38
39 >>> s.getvalue()
40
41 b‘binary data‘
42
43 >>>

readinto | bytearray | memoryview

 1 record_size = 32
 2
 3 buf = bytearray(record_size)
 4
 5 with open(‘somefile‘, ‘rb‘) as f:
 6
 7 while True:
 8
 9 n = f.readinto(buf)
10
11 if n < record_size:
12
13 break
14
15
16
17 另外有一个有趣特性就是 memoryview ，它可以通过零复制的方式对已存在的缓冲
18
19 区执行切片操作，甚至还能修改它的内容。比如：
20
21 >>> buf
22
23 bytearray(b‘Hello World‘)
24
25 >>> m1 = memoryview(buf)
26
27 >>> m2 = m1[-5:]
28
29 >>> m2
30
31 <memory at 0x100681390>
32
33 >>> m2[:] = b‘WORLD‘
34
35 >>> buf
36
37 bytearray(b‘Hello WORLD‘)

mmap

-- 内存映射的二进制文件

import os

import mmap

def memory_map(filename , access=mmap.ACCESS_WRITE):

size = os.path.getsize(filename)

fd = os.open(filename , os.O_RDWR)

return mmap.mmap(fd ,size , access=access)

size = 1000000

with open(‘data‘ , ‘wb‘) as f:

f.seek(size - 1)

f.write(b‘\x00‘)

m = memory_map(‘data‘)

print(len(m))

print(m[0:10])

# print(b‘hello daming‘ , file=m[0:11])

# m.close()

with open(‘data‘ ,‘rb‘) as f:

print(f.read(11))

os.path

-- basename | dirname | join | splitext | expanduser

 1 import os
 2
 3 >>> path = ‘demo.bin‘
 4
 5 >>> os.path.basename(path) #文件名
 6
 7 ‘demo.bin‘
 8
 9 >>> os.path.dirname(path) #文件路径
10
11 ‘‘
12
13 >>> os.path.join(‘tmp‘,‘data‘ , os.path.basename(path))  #路径拼接
14
15 ‘tmp\\data\\demo.bin‘
16
17 >>> path = ‘~/data/data.bin‘
18
19 >>> os.path.expanduser(path) #系统路径
20
21 ‘C:\\Users\\Administrator/data/data.bin‘
22
23 >>> os.path.splitext(path) #获取后缀名
24
25 (‘~/data/data‘, ‘.bin‘)

-- exists | isdir | islink | realpath

os.path.exists(‘/etc/passwd‘)   #判断文件是否存在

>>> os.path.isdir(‘user.data‘)  #判断是否是目录

False

>>> os.path.islink(‘demo.bin‘)  #判断是否是超链接

False

>>> os.path.realpath(‘demo.bin‘) #获取文件的绝对路径

‘E:\\zeopean\\pycode\\pyfunc\\demo.bin‘

-- getmtime | getsize 获取元数据

 1 >>> os.path.getmtime(‘demo.txt‘)
 2
 3 1459387735.7783203
 4
 5 >>> import time
 6
 7 >>> time.ctime(os.path.getmtime(‘demo.bin‘))
 8
 9 ‘Thu Mar 31 10:27:36 2016‘
10
11 >>> os.path.getsize(‘demo.txt‘)
12
13 47

-- listdir

>>> names = os.listdir(‘.‘)

>>> names

[‘closure.func.py‘, ‘closureInstance.py‘, ‘data‘, ‘demo.bin‘, ‘demo.txt‘, ‘file.

txt‘, ‘make_element.py‘, ‘mmap.demo.py‘, ‘readinto.demo.py‘, ‘yield.func.py‘]

tempfile

 1 --临时文件 TemporaryFile
 2
 3 from tempfile import TemporaryFile
 4
 5 f = TemporaryFile(‘w+t‘)
 6
 7 # Use the temporary file
 8
 9 ...
10
11 f.close()
12
13
14
15 --有名字的临时文件 NamedTemporaryFile
16
17 from tempfile import NamedTemporaryFile
18
19 with NamedTemporaryFile(‘w+t‘) as f:
20
21 print(‘filename is:‘, f.name)
22
23
24
25 -- 临时目录 TemporaryDirectory
26
27 from tempfile import TemporaryDirectory
28
29 with TemporaryDirectory() as dirname:
30
31 print(‘dirname is:‘, dirname)
32
33
34
35 -- 插件临时文件 mkstemp | mkdtemp
36
37 >>> import tempfile
38
39 >>> tempfile.mkstemp()
40
41 (3, ‘C:\\Users\\ADMINI~1\\AppData\\Local\\Temp\\tmp_jbpp74i‘)
42
43 >>> tempfile.mkdtemp()
44
45 ‘C:\\Users\\ADMINI~1\\AppData\\Local\\Temp\\tmpkwsqedpa‘
46
47
48
49 -- 获取临时文件名  gettempdir
50
51 import tempfile
52
53 tempfile.gettempdir()
54
55 ‘C:\\Users\\ADMINI~1\\AppData\\Local\\Temp‘

Pickle

 1 -- 序列化对象
 2
 3 >>> import pickle
 4
 5 >>> f = open(‘somedata‘, ‘wb‘)
 6
 7 >>> pickle.dump([1, 2, 3, 4], f)
 8
 9 >>> pickle.dump(‘hello‘, f)
10
11 >>> pickle.dump({‘Apple‘, ‘Pear‘, ‘Banana‘}, f)
12
13 >>> f.close()
14
15 >>> f = open(‘somedata‘, ‘rb‘)
16
17 >>> pickle.load(f)
18
19 [1, 2, 3, 4]
20
21 >>> pickle.load(f)
22
23 ‘hello‘
24
25 >>> pickle.load(f)
26
27 {‘Apple‘, ‘Pear‘, ‘Banana‘}