Get some elements from numpy array in reverse order

Question

I need to generate numpy arrays getting elements in reverse order from another array.

Toy example code

Lets say I use following toy example code:

import numpy as np
a = np.array([1,2,3,5,8,13])
n = len(a)

for i in range(n):
    print a[n:n-i-2:-1]

I would expect that last print is a [13 8 5 3 2 1], however I get an empty array [] as seen below:

>>> 
[13]
[13  8]
[13  8  5]
[13  8  5  3]
[13  8  5  3  2]
[]

Fix

So I had to create below fix to my code within the for loop to get what I expect.

for i in range(n):
    print a[n-i-1:n+1][::-1]

Which means selecting the arrays from the original array and then reversing it.

Questions

Why when I try a[6:0:-1] I get [13, 8, 5, 3, 2] but once I try a[6:-1:-1] I get an emtpy array []? I would expect to get the whole array reversed as when you try a[::-1].
Is the fix I implemented the way to go or there is something I'm missing here?

Edit

The post Understanding slice notation answers my first question but not the second one.

You might find this helpful https://stackoverflow.com/questions/509211/understanding-slice-notation — yatu, Jun 18 '19 at 13:02
Possible duplicate of [Understanding slice notation](https://stackoverflow.com/questions/509211/understanding-slice-notation) — WhySoSerious, Jun 18 '19 at 13:05
Sorry @ScottHunter I misread you question `a[6:5:-1]` would produce [] — Cedric Zoppolo, Jun 18 '19 at 13:24
And -1 is a "shorthand" for the index of the last element (in this case, 5), so it should (and does) do the same. — Scott Hunter, Jun 18 '19 at 13:26

score 3 · Accepted Answer · edited Nov 18 '22 at 20:45

3

Here is a slightly more elegant fix.

for i in range(n):
    print(a[n:-i-2:-1])

or even more elegant

for i in range(n):
    print(a[:-i-2:-1])

both print

[13]
[13  8]
[13  8  5]
[13  8  5  3]
[13  8  5  3  2]
[13  8  5  3  2  1]

It also demonstrates the rule that negative indices count from the top, which is why your original loop switches behavior when the stop index gets to -1.

edited Nov 18 '22 at 20:45

Shovalt

6,407
2
36
51

answered Jun 18 '19 at 13:28

Paul Panzer

51,835
3
54
99

This makes sense now. – Cedric Zoppolo Jun 18 '19 at 13:41

score 1 · Answer 2 · answered Jun 18 '19 at 13:31

1

Another option would be to use below code:

import numpy as np
a = np.array([1,2,3,5,8,13])
n = len(a)

l = range(n)
l[n-1] = None

for i in range(n):
    print a[n:l[n-i-2]:-1]

As performing a[n:None:-1] is same as a[::-1]

answered Jun 18 '19 at 13:31

Cedric Zoppolo

4,271
6
29
59

seralouk · Answer 3 · 2019-06-18T13:39:00.523

1

Use list comprehension and get the results in one line:

import numpy as np

a = np.array([1,2,3,5,8,13])
n = len(a)

# use list comprehension here
[list(a[n:-i-2:-1]) for i in range(n)]

# [[13],
#  [13, 8],
#  [13, 8, 5],
#  [13, 8, 5, 3],
#  [13, 8, 5, 3, 2],
#  [13, 8, 5, 3, 2, 1]]

In case you really need the exclicit for loop use this:

for i in range(n):
    print(a[n:-i-2:-1])

edited Jun 18 '19 at 13:39

answered Jun 18 '19 at 13:33

seralouk

30,938
9
118
133

Get some elements from numpy array in reverse order

Toy example code

Fix

Questions

Edit

3 Answers3