Typing list comprehensions (Py 3.9 and Py 3.10)

[dineshbvadhia]

What is the correct way to type list comprehensions? For example:

my_numbers = [1, 2, 3, 4, 5, 6, 7, 8, 9]

my_squares = [x + sqr(x) for x in my_numbers]

[srittau]

The result of a list comprehension is simply a list. Often the type can be determined automatically by a type checker, but an explicit annotation is also possible:

my_numbers: list[int] = [1, 2, 3, 4, 5, 6, 7, 8, 9]
my_squares: list[float] = [x + sqr(x) for x in my_numbers]

This requires from __future__ import annotations in Python 3.9.

[sobolevn]

I would add that annotations here are completely optional:

my_numbers = [1, 2, 3, 4, 5, 6, 7, 8, 9]
reveal_type(my_numbers)  # N: Revealed type is "builtins.list[builtins.int*]"

my_squares = [x + x for x in my_numbers]
reveal_type(my_squares)  # N: Revealed type is "builtins.list[builtins.int*]"

Most type-checkers will just infer the correct type.

[dineshbvadhia]

I was asking about the x in the list comprehension?

[srittau]

The x can't be annotated directly. If my_numbers is annotated correctly or the type is inferred correctly (you can check that with reveal_type), x should also be inferred correctly. For example, if my_numbers is list[int], type checkers infer correctly that x is int.