Python Pickle

Pickle is a standard Python module for serializing and deserializing Python objects. It is used to save objects to a file and load them back later.

Warning

Pickle is not secure. Never unpickle data received from an untrusted or unauthenticated source.

Pickle vs JSON

• Pickle is a binary format, while JSON is a text format
• Pickle is more efficient than JSON
• Pickle is not human-readable
• Pickle is not secure (can execute arbitrary code)

Example

import pickle

class Person:
    def __init__(self, name, age):
        self.name = name
        self.age = age

    def __repr__(self):
        return f'Person({self.name}, {self.age})'

person = Person('John', 30)

Save the object to a file:

with open('person.pickle', 'wb') as f:
    pickle.dump(person, f)

Load the object from a file:

with open('person.pickle', 'rb') as f:
    person = pickle.load(f)

Tip

Use cPickle instead of pickle for better performance (C implementation up to 1000 times faster)

Multiple Objects

import pickle

p1 = Person('John', 30)
p2 = Person('Jane', 25)

with open('people.pickle', 'wb') as f:
    pickle.dump(p1, f)
    pickle.dump(p2, f)

Load the objects from a file in the same order:

with open('people.pickle', 'rb') as f:
    p1 = pickle.load(f)
    p2 = pickle.load(f)

Pickle DataFrames

picklespandas

data = {}
data['df1'] = df1
data['df2'] = df2

with open('data.pickle', 'wb') as f:
    pickle.dump(data, f)

with open('data.pickle', 'rb') as f:
    data = pickle.load(f)

data1 = data['df1']
data2 = data['df2']

df1.to_pickle('df1.pickle')
df2.to_pickle('df2.pickle')

df1 = pd.read_pickle('df1.pickle')
df2 = pd.read_pickle('df2.pickle')

---