Ignalion's twist clear solution for Flatten a List by nickie - python coding challenges

Enable Javascript in your browser and then refresh this page, for a much enhanced experience.

Ignalion's twist solution in Clear category for Flatten a List by nickie

def flat_list(l):
    r = []
    def f(l):
        for i in l:
            r.append(i) if type(i) is int else f(i)
    f(l)
    return r

April 24, 2014

Comments:

veky on April 25, 2014, 8:35 p.m.
LOL. Nice closure. :-D

artemrudenko on Aug. 18, 2014, 5:13 p.m.
The only thing to add here is that it's recomended to use isinstance instead of type.

tommycarstensen on April 25, 2015, 10:24 p.m.
@artemrudenko Why is it recommended to use isinstance instead of type?

artemrudenko on April 28, 2015, 4:46 a.m.
Hi, For example from: <a href="https://stereochro.me/ideas/type-vs-isinstance">https://stereochro.me/ideas/type-vs-isinstance</a> And: <a href="http://stackoverflow.com/questions/1549801/differences-between-isinstance-and-type-in-python">http://stackoverflow.com/questions/1549801/differences-between-isinstance-and-type-in-python</a>

tommycarstensen on April 28, 2015, 7:25 a.m.
Thanks, that was good reading. It makes me wonder, why type even exists?

veky on July 11, 2015, 7:45 p.m.
Because something has to be at the bottom of the hierarchy. Python is too Western to be comfortable with "subclasses all the way down" approach. :-D And sometimes you need finer distinction. Not every inheritance is just for interfaces. Just recently, I had to write a class that's very similar to another I wrote before. Of course, I used inheritance. <pre class='brush: python'>class declared_foreign(foreign): </pre> But later, when time came for resolving foreign objects, I had <pre class='brush: python'>if isinstance(value, foreign): ... = value.resolve() </pre> and I could just add <pre class='brush: python'>elif isinstance(value, declared_foreign): ... </pre> but it wouldn't work (do you see why?:). Of course, I could switch the branches and then it would work, but that's really no way to write maintainable code. On the other hand, <pre class='brush: python'>if type(value) is foreign: ... = value.resolve() elif type(value) is declared_foreign: ... </pre> is clear and not perfidly waiting to introduce subtle bugs. <hr /> Also, type is extremely useful as a fixed point of itself <pre class='brush: python'>type(type) is type </pre> which makes it usable as an initial metaclass: <pre class='brush: python'>class C(int): x = 5 def f(self): return self * 3 </pre> is executed as <pre class='brush: python'>C = type('C', (int,), dict(x=5, f=lambda self: self*3)) </pre> But that's probably over your head right now.

nickie on April 28, 2015, 8:34 p.m.
In spite of all the good discussion above and the validity of the arguments in the two links (yes, `type` will not do you any good if you have a non-trivial, i.e., non-flat, hierarchy of objects), I'm arguing that its use is perfectly reasonable for this task. The task description reads: "There is a list which contains integers..." In my mind, an "integer" is an object of type `int`. Now, of course, in Python, one may define, e.g. <pre class='brush: python'>class my_int(int): pass </pre> then be able to compute `foo(6)*foo(7)` (which surprisingly evaluates to 42 of type `int`) and, of course have a list such as `[foo(1), foo(2), foo(3)]`. Is such a list considered to be a "list of integers"? My program, using `type`, answers no. (In fact, my program fails on such an input unless `foo` is iterable; I admit this behaviour as I consider the input to be invalid.) A different program, using `instanceof`, would answer yes and would probably flatten correctly the strange list.

veky on July 11, 2015, 7:57 p.m.
First, there is a subclass of int already existing. <pre class='brush: python'>>>> int.__subclasses__() [<class 'bool'>] </pre> Is `[5, True]` a list of integers? :-D Of course, it is a sophistic question. But I think you might be interested that "a strict subclass of int" is not so strange a beast in regular Python. Second: <blockquote> which surprisingly evaluates to 42 of type int </blockquote> And what did you expect? That it evaluates to foo(42)? :-) Technically, not every type is usable as a cast. (Think `datetime.date`.) Semantically, there's no reason to expect foo is homomorphic. (Think foo(x) as being "x with an error of 1".) Or did you expect `TypeError: unsupported operand type(s) for +: 'foo' and 'foo'`? That would be a plain lie. `foo(6)` is a number 6, as far as the int part of its personality is concerned. Of course it knows how to multiply with number 7. The whole point of inheritance is overriding only what you need, and inheriting the rest.

nickie on July 13, 2015, 7:48 a.m.
You're right, of course. As I said, my solution considers `[5, True]` not to be a list of integers. I think this is reasonable for the scope of this task.

viktor.pecheniuk on Dec. 1, 2014, 3:13 p.m.
nice:)

JohnOZ on May 1, 2015, 12:19 a.m.
Wow. My solution is exactly the same!

VincentDog on Nov. 10, 2020, 4:59 p.m.
true story

hihipch on Aug. 7, 2020, 9:07 a.m.
brilliant for making a recursive code in a short way

shm9688 on March 1, 2021, 9:13 p.m.
Wow,Understandable and so much to learn!

viktor.chyrkin on Jan. 23, 2022, 3:06 p.m.
recursive method

jang.cw on Nov. 17, 2022, 12:45 a.m.
cool and fast!

List, list and list again

0 %

You ClassRoom Demo is Ready With the new tool CheckiO ClassRoom, you will be able to analyze your students' progress, have your own leaderboard, change missions order on the map, and so on.