How to annotate mutable structures for Mypy

Question

I have this code:

def my_foo(x: dict[str, int | float], some_condition: bool) -> None:
    if some_condition:
        x['the_int'] = 1
    else:
        x['the_float'] = 1.0

my_dict = {'some_int': 2}

my_foo(my_dict, True)

Mypy (understandably) complains

error: Argument 1 to "my_foo" has incompatible type "dict[str, int]"; expected "dict[str, int | float]" [arg-type] note: "Dict" is invariant -- see https://mypy.readthedocs.io/en/stable/common_issues.html#variance src\type_test.py:23: note: Consider using "Mapping" instead, which is covariant in the value type

I tried do type-hint using MutableMapping but to no avail:

error: Argument 1 to "my_foo" has incompatible type "dict[str, int]"; expected "MutableMapping[str, int | float]"

This question is very common. The solution is to specify the type explicitly: my_dict: dict[str, int | float] = {'some_int': 2}. There should be a good dupe somewhere. — InSync
– InSync, Commented Oct 5, 2024 at 21:28
Slightly besides the point, but I think int | float is equivalent to float (you can pass an int anywhere a float is expected, just not vice versa) so you could simplify the annotation a bit by making it dict[str, float]. You do still need to explicitly annotate my_dict either way. — Samwise
– Samwise, Commented Oct 8, 2024 at 16:20
@Samwise Not quite, since extremely large integers don't have floating-point equivalents. — chepner
– chepner, Commented Oct 16, 2024 at 12:56

Peter Mortensen · Accepted Answer · 2025-05-31 14:04:29Z

3

D.lola's answer is preferable if you have a fixed mapping between what keys are ints and which are float. However, if you don't, you can also just tell mypy that my_dict can contain floats:

my_dict: dict[str, int | float] = {'some_int': 2}

mypy doesn't complain about this code (since it has special casing for literals) and then wont complain for further code.

edited May 31 at 14:04

Peter Mortensen

31.4k22 gold badges110 silver badges134 bronze badges

answered Oct 5, 2024 at 18:20

MegaIng

7,9442 gold badges24 silver badges39 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

D.lola Over a year ago

precisely, this works too, I didn't want to assume; plus x["the_int"] and x["the_float"] made me infer.

D.lola · Accepted Answer · 2025-05-31 20:58:54Z

3

You can try this. Create your own dict using TypedDict. You can use dataclass as well.

This code snippet should resolve the issue.

from typing import TypedDict, Union

class MyDict(TypedDict, total=False):
    some_int: int
    the_int: int
    the_float: float


def my_foo(x: MyDict, some_condition: bool) -> None:
    if some_condition:
        x["the_int"] = 1
    else:
        x["the_float"] = 1.0


my_dict: MyDict = {"some_int": 2}

my_foo(my_dict, True)

Tested result

mypy ol.py

Output:

Success: no issue found in 1 source file

edited May 31 at 20:58

answered Oct 5, 2024 at 18:16

D.lola

2,3222 gold badges9 silver badges22 bronze badges

Collectives™ on Stack Overflow

How to annotate mutable structures for Mypy

2 Answers 2

1 Comment

Tested result

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Tested result

Comments

Your Answer

Sign up or log in

Post as a guest

Related