Unexpected content of list comprehension with lambdas

Question

Hello SL community,
I would appreciate it if you help me with the following problem:
Can you explain the following code ?
https://code.sololearn.com/c9an6szV8etn/?ref=app

Answer 1

Kishalaya Saha yes...
a list comprehension generates a closure but not
a.... tuple comprehension...
i. E generator.

Answer 2

https://www.sololearn.com/Discuss/2049859/?ref=app

Answer 3

Ha, nice, I knew I had explained it somewhere but didn't know when and where. You guys found it! :) Thx, Qasem!

Answer 4

HonFu but not the function makes the closure.
Otherwise my linked prog couldn't work.
It is the list comprehension

Answer 5

yessine HonFu .....
It is worth that. Real cool stuff.

Answer 6

Oma Falk, ah, yes, I think, there's not really a difference in what we're saying.

In that other thread I just 'translated' the list comp to a regular nested function creating the same closure effect, in order to illustrate it.

Answer 7

HonFu
if you await behavior in that prog, you are
right

https://code.sololearn.com/cNU52yK69BSj/?ref=app

Answer 8

actually the second version is a generator, not a tuple , am I not right ?

Answer 9

Yeah, a generator is only a 'construction plan', so that generator will generate each function object as needed and hand over the latest i.

Interesting!

If I apply that old pattern of explanation, it would look like this now:

def cm():
def f(x):
return x*i
l = []
for i in range(3):
yield f

Answer 10

Katie Hattab you have a contribution to that topic ?

Answer 11

Yes, Qasem you found it, thank you

Answer 12

yessine you're absolutely right! It's a generator, not a tuple. And that explains it!

Answer 13

I try to explain easier:
All lambdas in cm depend on one variable i.
But one variable can only have one value - that is the last one assigned to it = 2.
Closures are really a bit tricky

Answer 14

In cm the function objects are not really built in list comprehension but only their description.
Although the defining function is gone, they remember life in their function and share the same last memory concerning i which is 2. So we have a closure.

Answer 15

cm2 is not a closure all lambda are independent of their birthplace and share no common value.

Answer 16

By the way...it is easy to curate since tuples don't create an own closure
https://code.sololearn.com/cJiE17h7fR1v/?ref=app

Answer 17

Oma Falk, seems I have to catch up with what you wrote and think about it for a moment...

Answer 18

Oma Falk I think difference in your version might have something to do with tuples being immutable. If we replace it by a list, a set, or even a dictionary, it works (fails?) like before.

P.S. I, too, attempted to explain this code nearly a year ago 😄 It's pretty similar to what you and HonFu said.
https://www.sololearn.com/Discuss/1577036/list-lambda

Edit: turns out it's not a tuple, but a generator, as pointed out later. And it makes a lot more sense.

Answer 19

Oma Falk I don't know what's a function description.. I guess I'll have to learn about it.