Weird lambda behaviour in list comprehension
Question:
I’m playing with lambda functions inside of list comprehension, and found some weird behaviour
x = [(lambda x: i) for i in range(3)]
print(x[0](0)) #print 2 instead of 0
print(x[1](0)) #print 2 instead of 1
print(x[2](0)) #print 2
Can someone explain why the result is not that I expect?
Answers:
lambdas bind variables themselves, not the values that they had. i is changed to 2 at the end of the list comprehension, so all the lambdas refer to i at that point, and thus refer to 2.
To avoid this, you can use the default argument trick:
[lambda x,i=i:i for i in range(3)]
This binds the value of i in the default argument (which is evaluated at function definition time).
I’m playing with lambda functions inside of list comprehension, and found some weird behaviour
x = [(lambda x: i) for i in range(3)]
print(x[0](0)) #print 2 instead of 0
print(x[1](0)) #print 2 instead of 1
print(x[2](0)) #print 2
Can someone explain why the result is not that I expect?
lambdas bind variables themselves, not the values that they had. i is changed to 2 at the end of the list comprehension, so all the lambdas refer to i at that point, and thus refer to 2.
To avoid this, you can use the default argument trick:
[lambda x,i=i:i for i in range(3)]
This binds the value of i in the default argument (which is evaluated at function definition time).