What is the different between these two list comprehension calculations?

Question

I've been running my own statistical analyses, but I'm curious as to why there are different outputs for the following two list comprehensions. I am using Jupyter Notebook on Ubuntu LTS 16.04:

Input:

example = [[15, 16, 17],
           [18, 19, 20]]

list1 = [item[0] + item[1] + item[2] for item in example]

for item in example:

    list2 = [item[0] + item[1] + item[2]]

list1, list2

Output: ([48, 57], [57])

Clearly the second function is printing the sum of the second row, but why not the first?

Answer 1

The second example is not a list comprehension, the final list has only one item as you're overwriting list2 variable with every loop execution.

Answer 2

To get the same output, the second loop should be corrected to this:

list2 = []
for item in example:
    list2.append(item[0] + item[1] + item[2])

Otherwise you overwrite what you had assigned to list2 in the previous iteration of the loop, only getting the result of the last iteration.