DiddiLeija's website

My personal website.

View on GitHub

Jumping from Python 2 to Python 3

Diddi Jumping from a "2" to a "3"

Overview

On January 1st, 2020, the life of the last release for Python 2 (2.7) ended in favor of Python 3. After that, many projects and packages dropped the support for it.

I’ve been working with the Pip project, and we are still receiving issues from Python 2. Some people even get annoyed because of weird errors, but we explain that their bugs are directly related to the Python 2 incompatibility. On this article, I will talk about the major differences between Python 2 and 3 (currently, the oldest non-EOL’ed Python, 3.7).

Some differences

The print statement -> print() function

On most of the Python 2 releases, you can print something by doing:

print "hello world!"

But in Python 3, that statement was replaced by the built-in function print():

print("hello world!")

raw_input() -> input()

On Python 2, there is a built-in function named raw_input(), to get user input:

name = raw_input("Enter your name: ")

But in Python 3, the function was renamed to input():

name = input("Enter your name: ")

Type hints

Python 3 (well, Python 3.5) introduced type hints (annotations):

def function_with_integers(arg: int) -> None:
    pass

function_with_integers(123)  # OK
function_with_integers(" ")  # wrong?

Other syntax changes

The with, for, if/elif/else, try/except/finally, def and class blocks had several changes and new features.

New modules

The Python 3 standard library has included a lot of new modules, and changed the existing ones. A great example is the typing module, created since Python 3.5. It provides type hint support, and many projects use it. But if you try to run code that uses typing on Python 2, it won’t work. That happens with many other popular modules, like urllib or http.

Default encoding for strings

On Python 2, all the strings are treated like bytes by default, even when Unicode strings are supported:

# try it with 3 strings
>>>s = "abcdef"
>>>b = b"abcdef"
>>>u = u"abcdef"
# now compare them
>>>u
u"abcdef"
>>>b
b"abcdef"
>>>s
b"abcdef"

But in Python 3, the strings are encoded by default:

# try it with 3 strings
>>>s = "abcdef"
>>>b = b"abcdef"
>>>u = u"abcdef"
# now compare them
>>>u
"abcdef"
>>>b
b"abcdef"
>>>s
"abcdef"

You don’t actually have to create u-strings!

How is that going to affect you?

As I were saying, many popular projects have finished the process of removing the Python 2 support (and btw, the support for later EOL Pythons). Others are working on it. Anyway, you won’t be able to use the latest versions of great packages if you still have an old Python.

A clear example is Pip, which is the Python package installer. Many users have found weird error messages when installing things with Pip:

ModuleNotFoundError: No module named 'typing'

NameError: Name 'input' is not defined

SyntaxError: invalid syntax

What do I have to do now?

The best thing you can do now, is forget Python 2 and directly jump into Python 3 (At this point, consider Python 3.7 to 3.10). This may require a great effort from you, but it will be the best.

Or, in the case that you need it, try not to get the newest releases of popular packages, because you will have trouble! However, this is not so efficient in the practice…

And this is where the article ends. I hope you like it, and maybe share this with more people. To go back to the main page of my website, press here.