Serializing Objects in Python

Serialization is the process of converting an object into a series of bytes, called a byte stream, to be stored or sent for later use. This could include being sent across a network, written to a file, or other process. In Python, this is called pickling.

Now, we’ve seen how to convert numbers to a string, so we can write them to a text file. But what about other, more complex, types of data, like a list, set, etc. Well, this is where serialization comes into play.

Serializing Data to Write to a File

There are several steps you’ll take to serialize your data:

1. Import the pickle module
2. Open the file for binary writing
3. Call the dump() method of the pickle module
4. Close the output file

Step one only needs to occur once in your application. Steps 2 through 4 may occur several times depending upon how your application is structured, and/or if more than one file needs to be written to.

Let’s look at an example where you do this:

import pickle
phonebook = {'James':'555-1234','Mary':'555-5678','John':'555-2468','Martha': '555-1357'}
out_file = open('phonebook.dat','wb')
pickle.dump(phonebook, out_file)
out_file.close()

Binary Files

You can see from the example, we open an out_file with ‘wb’. That opens a file for writing in binary format. This is different from ‘w’ which opens a file for writing a text file.

That file is a binary file. If you open it with a text editor (like Notepad++) you will view data that isn’t very readable that might look something like what you see below.

€}q (X   Jamesq
X   555-1234qX   MaryqX   555-5678qX   JohnqX   555-2468qX   MarthaqX   555-1357qu.

Reading Serialized Data (Deserialization)

Once you write a file, you will need to read that data back in. There is a similar process:

1. Import the pickle module
2. Open the file for binary reading
3. Import the data using the pickle.load method
4. Closing the data file.

Here is an example of that.

import pickle
in_file = open('phonebook.dat','rb')
phonebook = pickle.load(in_file)
in_file.close()
print(phonebook)

I will note here that pickling data is not considered secure. So you should only unpickle data that you get from a trusted source.

Another option is to use JSON (JavaScript Object Notation) which provides for a more secure method, and allows you to view the data since it comes in a special text format.

Serializing Objects in Python was originally found on Access 2 Learn