Pickling

The pickle module in Python is used for serializing and deserializing Python objects.

A byte stream is a sequence of bytes (numbers between 0 and 255) that represents raw binary data.

• represents the object’s structure in memory
• A text stream represents characters (using encodings like UTF-8)

• Serialization (or “pickling”) means converting a Python object into a byte stream (a format that can be stored in a file, transmitted over a network, etc.).
• Deserialization (or “unpickling”) is the reverse process of converting the byte stream back into a Python object.
• Pickling allows you to save complex data structures, such as lists, dictionaries, or even custom Python class instances, to a file or transfer them between different programs.

""" ===== Pickling (Serialization) ===== """
import pickle
 
# Example dictionary
data = {'Name': 'Alice', 'Age': 25, 'Country': 'USA'}
 
# Open a file to save the data
with open('data.pkl', 'wb') as file:
    # Serializes the data and saves it to file
    pickle.dump(data, file)  
 
 
"""Unpicklng (Deserialization) """
# Open the file and load the data back
with open('data.pkl', 'rb') as file:
    # Deserializes the byte stream back into a Python object
    loaded_data = pickle.load(file)  
 
print(loaded_data)
# {'Name': 'Alice', 'Age': 25, 'Country': 'USA'}

• In open(), we use wb because we’re writing it as binary and rb because we’re reading as binary
• file extensions don’t matter - you can use .pkl or .pickle
• dump → serialization
• load → deserialization

Using Pickling with JSON

• Python has a module named json so we can work with jsons
  • We can encode and decode from python to json and vise versa

import json

Common Use Cases

• Saving model states (e.g., machine learning models) to disk for later use.
• Storing large Python data structures for persistence.
• Transferring Python objects between different programs or systems.

Limitations

• The pickle module is Python-specific, meaning the pickled data can only be understood by Python programs.
• Pickling can be insecure if you unpickle data from untrusted sources, as it may execute arbitrary code.