List slice python

Question

I am attempting to quicksort a list and want to do so in place.

I have the following code and want to know in general (this problem comes up often for me as I try to solve list problems in place on the actual list rather than a complicated sequence of returns or using weird indices) can a list slice ever allow me to affect the list from which it's derived? Are there larger ramifications for doing something like this?

def partition(A,l,r):
    p=A[l]
    i=l+1
    print(A)
    for j in range(l+1,r):
        if A[j]<p:
            save=A[i]
            A[i]=A[j]
            A[j]=save
            i+=1
    save=A[i-1]
    A[i-1]=A[l]
    A[l]=save
    return i-1
def quickSort(A,n):
    if n>1:
        split=partition(A,0,n)
        quickSort(A[:split],len(A[:split]))
        quickSort(A[split+1:],len(A[split+1:]))
        print(A)
    return A
A=[3,4,5,2,10,7,6,9,1]
print(quickSort(A,len(A)))

A list slice is always a copy. That's why `l[:]` is the idiom for making a copy of a whole list. — Barmar, Feb 07 '17 at 10:15
So is there any way to get code like this to work? Can I actually point to the elements in my list A using a list slice somehow? — Chris L., Feb 07 '17 at 10:16
there's a disrepancy in the number of arguments of your `quickSort` function — Jean-François Fabre, Feb 07 '17 at 10:16
Possible duplicate of [Reference to Part of List - Python](http://stackoverflow.com/questions/1788608/reference-to-part-of-list-python) — Barmar, Feb 07 '17 at 10:17
A tip: idiomatic swapping in python without temporary variables: `A[i], A[j] = A[j], A[i]` — juanpa.arrivillaga, Feb 07 '17 at 10:22
Except I am not looking to copy my list in another variable, I literally want only a portion of it to go into the recursive call — Chris L., Feb 07 '17 at 10:22
I'm not sure what you mean exactly by "I literally want only a portion of it to go into the recursive call", slicing a list *always* returns a copy, so whatever you want, you cannot do that with slices. You will have to keep track of indices. — juanpa.arrivillaga, Feb 07 '17 at 10:27
look at the comment below...exactly the type of functionality I was looking for — Chris L., Feb 07 '17 at 10:28

hiro protagonist · Accepted Answer · 2017-02-07T10:38:06.087

4

numpy lets you assign to slices in a way that is reflected in the original array (numpy slices return views):

import numpy as np

a = np.array(range(10))
print(a)   # [0 1 2 3 4 5 6 7 8 9]
b = a[3:7]
print(b)   # [3 4 5 6]
b[0] = 33
print(b)   # [33  4  5  6]
print(a)   # [ 0  1  2 33  4  5  6  7  8  9]

a slice of a python list always returns a copy so you would have to write a view on a list yourself...

edited Feb 07 '17 at 10:38

answered Feb 07 '17 at 10:26

hiro protagonist

44,693
14
86
111

Man you are awesome! works perfectly and I would have never thought of using numpy for a task like this. Thank you thank you thank you. – Chris L. Feb 07 '17 at 10:42

List slice python

1 Answers1