Python Data Persistence – pickle Module
The serialization format used by the pickle module, which is a part of Python’s built-in module library, is very Python-specific. While this fact can work as an advantage that it doesn’t face any restrictions by certain external standards such as JSON format, the major disadvantage is that non-Python applications may not be able to reconstruct ‘pickled’ objects. Also, the pickle module is not considered secure when it comes to unpickling data received from an unauthenticated or untrusted source.
The pickle module defines the module-level dumps ( ) function to obtain a byte string ‘pickled’ representation of any Python object. Its counterpart function loads () reconstructs (‘unpickles’) the byte string identical Python object.
Following code snippet demonstrates the use of dumps () and loads () functions:
Example
>>> import pickle >>> numbers=[10,20,30,40] >>> pickledata=pickle.dumps(numbers) >>> pickledata b'\x80\x03]q\x00(K\nK\xl4K\xleK(e.' >>> #unpickled data . . . >>> unpickledata=pickle.loads(pickledata) >>> unpickledata [10, 20, 30, 40] >>>
There are dump ( ) and load () functions that respectively write pickled data persistently to a file-like object (which may be a disk file, a memory buffer object, or a network socket object) having binary and write ‘wb’ mode enabled, and reconstruct identical object from a file-like object having ‘rb’ permission.
Example
>>> #pickle to file
. . .
>>> import pickle
>>> numbers=[ 10 , 20 , 30 , 40 ]
>>> file=open ('numbers .dat' , ' wb ')
>>> pickle . dump (numbers, file)
>>> file . close ()
>>> #unpickle from file
. . .
>>> file=open {1 numbers . dat' , ' rb' )
>>> unpickledata=pickle . load (file)
>>> unpickledata
[10, 20, 30, 40]
>>>
Almost any type of Python object can be pickled. This includes built-in types, built-in, and user-defined functions, and objects of user-defined classes.
The pickle module also provides object-oriented API as a substitute for module-level dumps () /loads () and dump () /load () functions. The module has a pickier class whose object can invoke dump () or dumps () method to ‘pickle’ an object. Conversely, the unpicker class defines load () and loads () methods.
Following script has a person class whose objects are pickled in a file using pickier class. Original objects are obtained by the load () method of unpicker class.
Example
from pickle import Pickier, Unpickler
class User:
def__init__(self,name, email, pw):
self.name=name
self.email=email
self.pw=pw
def__str__(self):
return ('Name: { } email: { } password: {}'. \ format(self.name, self.email, self.pw))
user1=User('Raj an', '[email protected]', 'rajanl23')
user2=User('Sudheer', '[email protected]', 's 11')
print ('before pickling..')
print (user1)
print (user2)
file=open (' users . dat' , ' wb' )
Pickier (file) .dump (userl)
Pickier (file) .dump(user2)
file.close ()
file=open ( ' users . dat' , ' rb ' )
obj 1=Unpickler (file) . load ()
print ('unpickled objects')
print (obj1)
obj2=Unpickler (file) . load()
print (obj2)
Output:
E:\python37>python pick1-example.py before pickling. Name: Rajan email: rl23w-gmail.com password: rajanl23 Name: Sudheer email: s.lKwgmail.com password: s_ll unpickled objects Name: Rajan email: rl2 3wgmail.com password: rajanl23 Name: Sudheer email: s.llwgmail.com password: s_ll E:\python3 7 >