nested-min-max (PLW3301)

Derived from the Pylint linter.

Fix is sometimes available.

What it does

Checks for nested min and max calls.

Why is this bad?

Nested min and max calls can be flattened into a single call to improve readability.

Example

minimum = min(1, 2, min(3, 4, 5))
maximum = max(1, 2, max(3, 4, 5))
diff = maximum - minimum

Use instead:

minimum = min(1, 2, 3, 4, 5)
maximum = max(1, 2, 3, 4, 5)
diff = maximum - minimum

Known issues

The resulting code may be slower and use more memory, especially for nested iterables. For example, this code:

iterable = range(3)
min(1, min(iterable))

will be fixed to:

iterable = range(3)
min(1, *iterable)

At least on current versions of CPython, this allocates a collection for the whole iterable before calling min and could cause performance regressions, at least for large iterables.

Fix safety

This fix is always unsafe and may change the program's behavior for types without full equivalence relations, such as float comparisons involving NaN.

print(min(2.0, min(float("nan"), 1.0)))  # before fix: 2.0
print(min(2.0, float("nan"), 1.0))  # after fix: 1.0

print(max(1.0, max(float("nan"), 2.0)))  # before fix: 1.0
print(max(1.0, float("nan"), 2.0))  # after fix: 2.0

The fix will also remove any comments within the outer call.