Does typing make python faster?

An effortless way to document your code and boost your development speed

When a developer starts writing Python code (without type hints), an IDE, like PyCharm, provides robust auto complete support. But as the number of lines and files increases, we start to notice the auto complete isn’t helpful at all. You are not getting those auto complete suggestions as they were earlier. What’s happening here? The reason this is happening is because the more code we encapsulate in functions, the more unclear variable data types get. Hence, with no information about variable data types, how can a linter function?

To tackle this issue, Python incorporated PEP 484 into the Python 3.5 release. Through these hints, a programmer can now document what type of data is allowed in a variable.

In this article, I am going to define type hints, what effect they have on the interpreter’s performance, and how they can help in documenting the code.

Statically typed vs Dynamically typed languages

Primarily, there are two types of programming languages. First are the statically typed languages such as C++ and Java, and second are the dynamically typed languages like JavaScript and Rust. Statically typed languages require variables to be declared with their data types before using them. On the contrary, dynamically typed languages work with variables declared during runtime. There is always a risk of encountering a runtime error in dynamically typed languages, so developers must take care of such scenarios by either using try-except or ensuring the return types.

In this Python code, if a number doesn’t match with any logic, then by default None is returned. When this happens, a runtime AttributeError is raised because we are trying to use .lower(), a string attribute, on the None data type.

Will using type hints make Python a statically typed language?

No, type hints don’t force Python to check the type of data being stored in a variable. They are just for the developer’s reference and help auto-completion software. They allow IDEs to recognise data types and give suggestions based on this information. It also serves a purpose of documenting your code in a way that other developers need not read the entire snippet to know a variable’s data type.

This gist explains how type hints are essentially ignored by Python interpreter.

How to write type hints?

The syntax goes like this

:  = 

Type hints not only validate data types but can also be used with a class name to indicate a class instance.

For instance, here the type hint provided is for class User, which has exist, reset, and validate_password methods declared, and we can see the linter already has those in suggestions for us. Similarly, a variable hinted as str will have all string related attributes in suggestion like lower, upper, title, etc. When types are specified for function parameters and a wrong data type variable is passed, then the developer is warned about a potential error by the linter.

Specifying function return types

In statically typed languages, functions must declare the type of data they are going to return. For example, in C++, a function that returns an integer would look like this.

Whereas in Python there is no such convention, a function can return any data type. This makes code difficult to debug and ambiguous. In Python, we do it like this, using a hyphen and greater than after closing parenthesis:

Reiterating the previous point, function return type hints do not force functions to return specific data types; they are used as a means of documentation and readability.

Does using type hints affect interpreter’s performance?

There’s no effect on performance whatsoever with or without type hints. Cpython ignore type hints similar to how comments are discarded during runtime.

Conclusion

To summarize the article, type hints:

1. Prevent us from making silly mistakes in attribute names.
2. Make code more readable.
3. Self documents the code.
4. Helps in auto completion.
5. Along with linter, warns developer of potential incompatibilities.

Thanks for reading till the end! If you liked this article or are interested in topics like Automation and Python, follow me or subscribe to my newsletter.

More content at plainenglish.io