如何用列表的副本初始化数据类字段？

Question

I'm parsing text files which have the following line structure:

name_id, width, [others]

After parsing an input, I'm feeding each parsed line into the dataclass :

@dataclass(frozen=True, eq=True)
class Node():
    name: str
    rate: int
    other: list[str] = field(compare = False)

In the middle of the code, I'm removing values from other , which results in removing values from original parse output. I don't want that. How can I make dataclass to force making a copy of the list?

In common class, as I understand, solution would be:

self.other = other.copy()

But if doing so, I should rewrite whole init method which breaks the purpose of dataclass .

So how to initialize dataclass field with a copy of a list?

I've tried doing this:

def __post_init__(self):
    setattr(self, 'other', self.other.copy())

Which yield can't assign attribute of frozen instance.

Answer 1

IMHO this is a bad idea and looks like breaking "explicit is better than implicit" aphorisms . But if you really want to you can hack force it with object.__setattr__ :

@dataclass(frozen=True, eq=True)
class Node:
    name: str
    rate: int
    other: list[str] = field(compare=False)

    def __post_init__(self):
        object.__setattr__(self, 'other', self.other.copy())

Explanation

object.__setattr__ is the default python __setatter__ while __setattr__ uses class's __setattr__ method which in the case of dataclass is overridden to ensure that user doesn't change the field of the frozen dataclass (so you skip the check).