sorting np array vs python list yields different results

Question

I wanted to do a little numpy practice, so I decided to borrow some quicksort code and change it to use numpy.

This is the original:

def quicksort(arr):
    if len(arr) <= 1:
        return arr
    pivot = arr[len(arr) // 2]

    left = [x for x in arr if x < pivot]
    middle = [x for x in arr if x == pivot]
    right = [x for x in arr if x > pivot]

    return quicksort(left) + middle + quicksort(right)

print (quicksort([1,2,32,121,4,5,51]))

And here I change it to use numpy arrays

def quicksort_np(arr):
    if len(arr) <= 1:
        return arr
    pivot = arr[len(arr) // 2]

    left = arr[arr < pivot]
    middle = arr[arr == pivot]
    right = arr[arr > pivot]

    return quicksort_np(left) + middle + quicksort_np(right)

print (quicksort_np(np.array([1,2,32,121,4,5,51])))

Running the original version, I get the correct output:

[1, 2, 4, 5, 32, 51, 121]

However, running the numpy version, I get an empty list:

[]

I can't understand what the problem is. I thought it had to do with the new lists not being copies, so I tried calling .copy() on each list inside quicksort_np, but I still got the same output.

Porting list code to numpy is not a good way to learn numpy. Superficial similarities can lead to bugs like this, and also promote poor numpy style. — hpaulj
– hpaulj, Commented Jan 23, 2018 at 3:56

smac89 · Accepted Answer · 2018-01-23 03:32:13Z

2

I think the error is in your line:

return quicksort_np(left) + middle + quicksort_np(right)

The '+' operator is a concatenation operator for python lists, but a sum operator for numpy arrays.

You can probably solve the problem by using np.concatenate

edited Jan 23, 2018 at 3:32

smac89

43.8k15 gold badges150 silver badges202 bronze badges

answered Jan 23, 2018 at 3:21

rickyalbert

2,6824 gold badges25 silver badges31 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

smac89 Over a year ago

Thanks very much. I guess I have to be careful of what I assume is common across python and numpy

jpp · Accepted Answer · 2018-01-23 03:23:20Z

2

The + operator works on python lists but not on numpy arrays.

Try np.hstack instead:

def quicksort_np(arr):
    if len(arr) <= 1:
        return arr
    pivot = arr[len(arr) // 2]

    left = arr[arr < pivot]
    middle = arr[arr == pivot]
    right = arr[arr > pivot]

    return np.hstack((quicksort_np(left), middle, quicksort_np(right)))

answered Jan 23, 2018 at 3:23

jpp

166k37 gold badges301 silver badges363 bronze badges

Collectives™ on Stack Overflow

sorting np array vs python list yields different results

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related