Why does a list of generators only return the elements of the last generator?
Question:
I’m given an arbitrary list of objects (e.g. ['foo', 'bar']
). My goal is to produce a list of equal size where every element in the result list is a generator that repeats the respective input element 5 times.
This is a big simplification of what I actually want to do, and I am aware that there are many ways in which to solve this task.
However, I have stumbled upon some weird behavior in how I wanted to solve this that I can not explain.
Here is my solution to the above task:
my_iterators = [
(element for _ in range(5))
for element in ["foo", "bar"]
]
for my_iterator in my_iterators:
print(list(my_iterator))
I now expected the output to be:
['foo', 'foo', 'foo', 'foo', 'foo']
['bar', 'bar', 'bar', 'bar', 'bar']
However, to my surprise it was instead:
['bar', 'bar', 'bar', 'bar', 'bar']
['bar', 'bar', 'bar', 'bar', 'bar']
Why is it that (element for _ in range(5))
seems to be an iterator over the last element in the input_list irregarding of what it actually is in the context of for element in ["foo", "bar"]
?
How would I need to adapt my code to produce what my original goal was?
Answers:
As pointed out by jonsharpe and David Buck in the comments, what’s actually causing the non-intuitive behavior here is Python’s late binding.
There seem to be multiple ways of fixing the posted code, one would be:
def make_iterator(element):
return (element for _ in range(5))
my_iterators = [
make_iterator(element)
for element in ["foo", "bar"]
]
for my_iterator in my_iterators:
print(list(my_iterator))
I’m given an arbitrary list of objects (e.g. ['foo', 'bar']
). My goal is to produce a list of equal size where every element in the result list is a generator that repeats the respective input element 5 times.
This is a big simplification of what I actually want to do, and I am aware that there are many ways in which to solve this task.
However, I have stumbled upon some weird behavior in how I wanted to solve this that I can not explain.
Here is my solution to the above task:
my_iterators = [
(element for _ in range(5))
for element in ["foo", "bar"]
]
for my_iterator in my_iterators:
print(list(my_iterator))
I now expected the output to be:
['foo', 'foo', 'foo', 'foo', 'foo']
['bar', 'bar', 'bar', 'bar', 'bar']
However, to my surprise it was instead:
['bar', 'bar', 'bar', 'bar', 'bar']
['bar', 'bar', 'bar', 'bar', 'bar']
Why is it that (element for _ in range(5))
seems to be an iterator over the last element in the input_list irregarding of what it actually is in the context of for element in ["foo", "bar"]
?
How would I need to adapt my code to produce what my original goal was?
As pointed out by jonsharpe and David Buck in the comments, what’s actually causing the non-intuitive behavior here is Python’s late binding.
There seem to be multiple ways of fixing the posted code, one would be:
def make_iterator(element):
return (element for _ in range(5))
my_iterators = [
make_iterator(element)
for element in ["foo", "bar"]
]
for my_iterator in my_iterators:
print(list(my_iterator))