On Mon, Apr 25, 2022 at 03:38:21PM -0000, aanonyme.personne@hotmail.fr wrote:
Typically, when subclassing a NamedTuple type, you often don't want the <, >, <=, >=, + or * operators to work, so in that case you would want for the related methods to return NotImplemented.
When I have subclassed NamedTuple types, I have never done that. If `obj` is a tuple, it supports those operators. If you subclass tuple, and get a NamedTuple, the subclass is still a tuple, and the Liskov Substitution Principle tells us that it should support all tuple operations. If you subclass the NamedTuple, that is still a tuple, and again Liskov tells us that it should behave like a tuple. I'm the first person to acknowledge that Liskov is more of a guideline than a law, so I won't say that what you are doing is *always* wrong, but surely it is wrong more often than it is right. In any case, your function is a one-line function. Not every one line function needs to be a builtin. If you don't want to reimplement it each and every time, it is easy enough to import it from your personal utility library: class Spam(MyNamedTuple): from mytoolbox import NotImplementedMethod __lt__ = __gt__ = NotImplementedMethod Its not quite as convenient as a builtin, but on the plus side, you don't have to write a PEP and then wait until Python 3.11 or 3.12 before you can start using it. -- Steve