4.6 The pickle Module

The pickle module, shown in Example 4-11, is used to serialize data—that is, convert data to and from character strings, so that they can be stored on file or sent over a network. It's a bit slower than marshal, but it can handle class instances, shared elements, and recursive data structures, among other things.

Example 4-11. Using the pickle Module

File: pickle-example-1.py

import pickle

value = (
    "this is a string",
    [1, 2, 3, 4],
    ("more tuples", 1.0, 2.3, 4.5),
    "this is yet another string"
    )

data = pickle.dumps(value)

# intermediate format
print type(data), len(data)

print "-"*50
print data
print "-"*50

print pickle.loads(data)

<type 'string'> 121
--------------------------------------------------
(S'this is a string'
p0
(lp1
I1
aI2
aI3
aI4
a(S'more tuples'
p2
F1.0
F2.3
F4.5
tp3
S'this is yet another string'
p4
tp5
.
--------------------------------------------------
('this is a string', [1, 2, 3, 4], ('more tuples',
1.0, 2.3, 4.5), 'this is yet another string')

On the other hand, pickle cannot handle code objects (but see the copy_reg module for a way to fix this).

By default, pickle uses a text-based format. You can also use a binary format, in which numbers and binary strings are stored in a compact binary format. The binary format usually results in smaller files. This is demonstrated in Example 4-12.

Example 4-12. Using the pickle Module in Binary Mode

File: pickle-example-2.py

import pickle
import math

value = (
    "this is a long string" * 100,
    [1.2345678, 2.3456789, 3.4567890] * 100
    )

# text mode
data = pickle.dumps(value)
print type(data), len(data), pickle.loads(data) == value

# binary mode
data = pickle.dumps(value, 1)
print type(data), len(data), pickle.loads(data) == value