为什么在这个例子中self和类名的用法不同?

我想你是在问 self.initial_account_id += 12 。这不会更新class属性;它创造了一个新的 例子 遮蔽类属性的属性,相当于 self.initial_account_id = self.initial_account_id + 12 对于每个新实例,右侧出现的 self.initial_account_id 尚不存在,因此移交给 BankAccount.initial_account_id .

假设您没有定义名为的实例属性 initial_account_id 那么 self.account_id = self.initial_account_id 将按预期工作(RHS决定 银行账户.initial_account_id ),但最好还是明确一点:

class BankAccount:


    initial_account_id = 1000110011
        
    def __init__(self, fname, lname, deposit = 0) -> None:
        
        BankAccount.initial_account_id += 12
        self.account_id = BankAccount.initial_account_id
        self.fname = fname
        self.lname = lname
        self.deposit = deposit

(也就是说,如果你 打算 要使用类属性,请这样说,而不是通过类的实例隐式访问它。)

更好的是,从中删除递增帐户ID的(显式)逻辑 BankAccount.__init__ ,并通过使用迭代器隐式地执行此操作。

from itertools import count


class BankAccount:
    initial_account_id = count(1000110011, 12)
        
    def __init__(self, fname, lname, deposit = 0) -> None:
        # As an aside, self.initial_account_id here would
        # be safe, since we aren't *assigning* to the attribute, only calling
        # one of its mutating methods.
        self.account_id = next(BankAccount.initial_account_id)
        self.fname = fname
        self.lname = lname
        self.deposit = deposit