Is there a reason why I should reverse these lines of code?

Question

I'm doing some practice on Dataquest.

Here's the exercise:

Create a function, sort, that:

Takes as input a list of lists.
Orders the outer list by the third element in the inner lists in descending order.
Resolves ties by sorting by the first one in ascending order.
Modifies the input in place..

What I've done:

def sort_index(some_list):
    some_list.sort(key=lambda x: x[2], reverse=True)
    some_list.sort(key=lambda x: x[0])

The output:

Function sort_index did not correctly modify the value of argument number 1.

Can someone explain to me why it worked when I reversed lines 1 and 2:

def sort_index(some_list):
    some_list.sort(key=lambda x: x[0])
    some_list.sort(key=lambda x: x[2], reverse=True)

If you change the order of the sorts, that changes the output — Samathingamajig, Aug 08 '22 at 20:34
@Samathingamajig That's obvious, the question is why the second order produces the correct output. — Barmar, Aug 08 '22 at 20:35

score 0 · Answer 1 · answered Aug 08 '22 at 20:34

The first .sort call is not needed, only the second one (with x[2]) matters if there are no ties. The only remaining behavior to understand is what happens to ties.

The list .sort method in Python is a stable sort, that's needed for it to work.

For example, if your input contains ('a', None, 5) and ('b', None, 5) tied on x[2], then the first .sort call will put them in this order (relative to each other) because 'a' < 'b', and the second .sort call will keep this order, because 5 == 5 and the sort is stable.

Is there a reason why I should reverse these lines of code?

1 Answers1