041 文件 持久化

pickle

• 序列化（持久化，落地）：把程序运行中的信息保存在磁盘上
• 反序列化： 序列化的逆过程
• pickle: Python 提供的序列化模块
  • pickle.dump: 序列化
  • pickle.load: 反序列化

举例

>>> import pickle  # 下面的例子就不重复这个导入操作了
>>> 

“腌制泡菜”

>>> age = 18
>>> with open("text.txt", "wb") as f:
...   pickle.dump(age, f)
>>> 

# text.txt
€K.

“取出泡菜”

>>> f = open("text.txt", "rb")  # 这里用二进制是必要的
>>> f.readline()
b'x80x03Kx12.'
>>> f.seek(0)
0
>>> pickle.load(f)
18
>>> 

序列化

>>> lst = [18, 19, 20, "york", "fish", [2, 3]]
>>> with open("text.txt", "wb") as f:
...   pickle.dump(lst, f)
...
>>> 

反序列化

>>> with open("text.txt", "rb") as f:
...   pickle.load(f)
...
[18, 19, 20, 'york', 'fish', [2, 3]]
>>> 

注意事项

1. pickle 几乎可以把所有 Python 的对象都转换为二进制的形式存放
2. 用 pickle 序列化，需要以二进制打开
3. 若要将 pickle 后的内容保存为文件，推荐以 .pkl 或 .pickle 为后缀

shelve

• 持久化工具
• 存取方式类似字典，用 k, v 保存数据

>>> import shelve  # 下面的例子就不重复这个导入操作了
>>> 

举例

>>> shv = shelve.open("shv.db")
>>> shv["one"] = 1
>>> shv["two"] = 2
>>> shv["three"] = 3
>>> shv.close()
>>> 

• Windows 下，上述操作过后，相应文件夹内出现三个文件
  • shv.db.bak
  • shv.db.dat
  • shv.db.dir

>>> shv = shelve.open("shv.db")
>>> try:
...   print(shv["one"])
...   print(shv["four"])
... except KeyError as e:
...   print(e)
... finally:
...   shv.close()
...
1
b'four'
>>> 

特性

• 不支持多个应用并行写入
  • 解决方法：open 的时候可以使用 flag=r
• 写回问题
  • shelv 默认情况下，不会接受持久化对象进行任何修改
  • 解决方法：强制写回 writeback=True

>>> shv = shelve.open("shv.db", flag='r')
>>> try:
...   k = shv["one"]
...   print(k)
... finally:
...   shv.close()
...
1
>>> 

>>> shv = shelve.open("shv.db")
>>> try:
...   shv["one"] = {'a':65, 'b':98, 'c':99}
... finally:
...   shv.close()
...
>>> shv = shelve.open("shv.db")
>>> try:
...   one = shv["one"]
...   print(one)
... finally:
...   shv.close()
...
{'a': 65, 'b': 98, 'c': 99}
>>> 

强制写回

问题

• shelve 关闭，内容仍存于内存中，没有写回数据库

>>> shv = shelve.open("shv.db")
>>> try:
...   k = shv["one"]
...   print(k)
...   k['a'] = 97
... finally:
...   shv.close()
...
{'a': 65, 'b': 98, 'c': 99}
>>> 

解法

>>> shv = shelve.open("shv.db", writeback=True)
>>> try:
...   k = shv["one"]
...   print(k)
...   k['a'] = 97
... finally:
...   shv.close()
...
{'a': 65, 'b': 98, 'c': 99}
>>> shv = shelve.open("shv.db")
>>> try:
...   print(shv["one"])
... finally:
...   shv.close()
...
{'a': 97, 'b': 98, 'c': 99}
>>> 

使用 with 管理上下文环境

>>> with shelve.open("shv.db", writeback=True) as shv:
...   print(shv["one"])
...   shv["one"]['a'] = 0
...
{'a': 97, 'b': 98, 'c': 99}
>>> with shelve.open("shv.db") as shv:
...   print(shv["one"])
...
{'a': 0, 'b': 98, 'c': 99}
>>>