Storing objects with the best-named module ever, pickle

I love pickle. I love that Python has a module named pickle. I have never used pickle before, but I like that it’s there. Let’s take a look at what pickle does.

Per the docs, pickle provides an algorithm for serializing and unserializing a Python object. When we pickle an object, we turn it into a stream of bytes that can be sent to another process or saved to a file to be reconstructed at a later time. You still have to define the class in the namespace of the process reading the pickle - a pickled object comes with object data but no way to tell the program about an object type that’s not already in the process namespace. pickle also comes with a faster cPickle companion that’s written in C to make it super fast.

We can kick things off like so:

import pickle

class Cucumber:

	def __init__(self, height):
		self.height = height

	def grow(self):
		self.height += 1
		return "Current height is: ", str(self.height)

veggie = Cucumber(2)

print pickle.dumps(veggie)


print pickle.dumps(veggie)

Here, I have created a Cucumber object type and given it a method grow(), which will increment its height. Calling the pickle.dumps() method on an instance of our Cucumber yields a short string of ASCII characters as a representation of our object.

I can also save objects to files and reload them in perfect condition:

veggie_two = Cucumber(4)

file_save = open('pickled_veggies', 'w')

pickle.dump(veggie_two, file_save)


file_read = open('pickled_veggies', 'r')

veggie_three = pickle.load(file_read)

print "Veggie Three Height: ", str(veggie_three.height)

pickle’s dump() mechanism (not dumps() with a ‘s’) is used to save these ASCII object strings to a file, and load() is used to read them back from that file. loads() would read them from a string.

And that’s what pickle does, simply and well. It makes your objects into strings, light, compact and ready to be stored forever.