What's an elegant way to create a list of objects a list of other objects in Python?-CodePudding

Imagine I have two dataclasses:

class A:
   x: int
   y: int

class B:
   x:int
   y:int

And I have a list of class a:

a_list = [A(x=1, y=2), A(x=3,y=4), ...]

What would be an elegant way of creating a list of class B from this list (copying the properties over in each case)?

CodePudding user response：

Give B a class method that produces an instance of B given an instance of A:

class B
    x: int
    y: int

    @classmethod
    def from_A(cls, a):
        return cls(a.x, a.y)

Then you can write

bs = [B.from_A(a) for a in a_list]

There's a bit of an asymmetry here: B knows details about A, but A knows nothing about B. You could reverse this and let instances of A produce instances of B:

class A:
   x: int
   y: int

   def to_B(self):
       return B(self.x, self.y)

bs = [a.to_B() for a in a_list]

Or, you could decide that neither A nor B should be "privileged" to know details about the other, instead having a third "omniscient" entity that knows how to go in one direction or the other. (This is at least symmetrical.)

def b_from_a(a):
    return B(a.x, a.y)

bs = [b_from_a(a) for a in a_list]

How you decide to encapsulate the creation of a B from an A is up to you; other factors in your code may help you decide which of the three options is most reasonable.

CodePudding user response：

Assuming both are dataclasses and all of the names match up, you can pass A as keyword arguments to B's constructor using asdict.

my_b_object = B(**asdict(my_a_object))

and, of course, to apply it over a list, use a list comprehension.

b_list = [B(**asdict(a)) for a in a_list]

CodePudding user response：

Access the attributes directly and feed them into B:

b_list = [B(a.x, a.y) for a in a_list]