Array into multiple arrays

Question

(Python) I have the following array

x=[1,2,3,4,5,6]

and I want this one

x=[[1,2],[2,3],[3,4],[4,5],[5,6]]

I used this function

def split(arr, size):
 arrs = []
 while len(arr) > size:
     pice = arr[:size]
     arrs.append(pice)
     arr   = arr[size:]
 arrs.append(arr)
 return arrs

But it only generates this

x=[[1,2],[3,4],[5,6]]

If looking for performance, use NumPy strides or `broadcasting` - https://stackoverflow.com/questions/40084931/taking-subarrays-from-numpy-array-with-given-stride-stepsize — Divakar, Apr 16 '18 at 17:48
You have a simple error. Change `arr = arr[size:]` to `arr = arr[1:]` — pault, Apr 16 '18 at 17:50
You say `x` is an array, but show a list. The code also generates lists. Is this a list problem or a numpy array one? I"m tempted to edit the tags, removing 'arrays', 'numpy' and 'split' — hpaulj, Apr 16 '18 at 17:55
@pault is almost correct. use `arr = arr[size-1:]` I assume you're passing 2 for size. You haven't defined the correct result for other values so it might not be correct in those cases. — Jeff Learman, Apr 16 '18 at 17:59
@JeffLearman I am pretty sure it should not be `size-1` although I agree that OP didn't specify. `arr = arr[1:]` returns a similar type output for all values of size (ie, lists of the same length, values consecutive). — pault, Apr 16 '18 at 18:03
@pault: Perhaps you're right: it depends on what `size` is intended to mean. `size-1` works when size is 2, and does *something* that might be meaningful when size is >2. — Jeff Learman, Apr 16 '18 at 18:08

rafaelc · Answer 1 · 2018-04-16T18:29:39.787

3

I suppose you want to develop your own code and not use libraries or built-in functions.

Your code is fine.

There's just one simple mistake: change the slice index from size to 1 in this line arr = arr[size:], where 1 means size - (size-1)

def split(arr, size):
    arrs = []
    while len(arr) > size:
        pice = arr[:size]
        arrs.append(pice)
        arr  = arr[1:]
    arrs.append(arr)
    return arrs

output:

[[1, 2], [2, 3], [3, 4], [4, 5], [5, 6]]

It also works for other sizes:

print split(x, 3)
print split(x, 4)

[[1, 2, 3], [2, 3, 4], [3, 4, 5], [4, 5, 6]]
[[1, 2, 3, 4], [2, 3, 4, 5], [3, 4, 5, 6]]

edited Apr 16 '18 at 18:29

answered Apr 16 '18 at 17:47

rafaelc

57,686
15
58
82

1

This doesn't work for `split(x, 3)`, etc. It *happens* to work in this one specific case, but it's wrong. – pault Apr 16 '18 at 17:48
1

you could simplify by noting that `size-(size-1) = 1` – pault Apr 16 '18 at 17:50
No I mean the correct thing to do is to not use `size` at all in the slice. The code should be `arr = arr[1:]`. You want to slide the list over by 1 each time. It is independent of `size`. – pault Apr 16 '18 at 17:53
@pault: Really? What's the semantics for `size`? The author doesn't say. – Jeff Learman Apr 16 '18 at 18:15
@JeffLearman this exchange happened before you posted your comment which correctly pointed out that OP didn't specify. I stand by my interpretation of the problem, though I recognize that you have a valid point. – pault Apr 16 '18 at 18:17

Sphinx · Answer 2 · 2018-04-16T18:06:18.333

loop your array with index, then put [index : index+size] as one element of new array.

The codes will be like:

x=[1,2,3,4,5,6]
size = 2
print([x[index:index+size] for index in range(0, len(x)-size+1)])
size = 4
print([x[index:index+size] for index in range(0, len(x)-size+1)])

Output:

[[1, 2], [2, 3], [3, 4], [4, 5], [5, 6]]
[[1, 2, 3, 4], [2, 3, 4, 5], [3, 4, 5, 6]]
[Finished in 0.189s]

or use zip() with size>1.

x=[1,2,3,4,5,6]
size = 2
print( [item for item in zip( *[x[index:len(x)+index-1] for index in range(0, size)])])
size = 4
print( [item for item in zip( *[x[index:len(x)+index-1] for index in range(0, size)])])

Output:

[(1, 2), (2, 3), (3, 4), (4, 5), (5, 6)]
[(1, 2, 3, 4), (2, 3, 4, 5), (3, 4, 5, 6)]
[Finished in 0.284s]

score 1 · Answer 3 · edited Apr 16 '18 at 17:58

1

Try this :

def split(arr, size):
    return [x[i:i+size] for i in range(0,len(x)-size+1)]

edited Apr 16 '18 at 17:58

rafaelc

57,686
15
58
82

answered Apr 16 '18 at 17:48

FadeoN

101
4

This could be right, depending on what's meant by `size`, which the OP doesn't specify. But it's very pythonic and shows the use of list comprehensions. And it's fixable if something else is meant by `size`. – Jeff Learman Apr 16 '18 at 18:17

Wasi · Answer 4 · 2018-04-16T17:55:02.050

1

You can use windowed() function from the more_itertools module.

from more_itertools import windowed
x = list(windowed(x, 2))

You can install it using pip

pip install more-itertools

edited Apr 16 '18 at 17:55

answered Apr 16 '18 at 17:49

Wasi

1,473
3
16
32

score 0 · Answer 5 · edited Apr 16 '18 at 17:51

0

a = [1,2,3,4,5]

answer = list(map(list, zip(a, a[1:])))

print(answer)

edited Apr 16 '18 at 17:51

rafaelc

57,686
15
58
82

answered Apr 16 '18 at 17:46

tkhurana96

919
7
25

I don't want it to be a tuple, but arrays inside an array – Beatriz Orozco Apr 16 '18 at 17:56
@BeatrizOrozco: this produces lists inside lists, which your example shows as the expected output. Tuples would be printed using parentheses, not square brackets. It would need to be modified for `size` other than 2. – Jeff Learman Apr 16 '18 at 18:19
I fixed it, thnk you – Beatriz Orozco Apr 16 '18 at 18:31

score 0 · Answer 6 · answered Apr 16 '18 at 18:40

You can use fancy indexing when using NumPy. This will require x to be converted to a numpy array. You will also need to create an array of incidences labeled ind below

x=[1,2,3,4,5,6]
x = np.array(x)

ind = np.arange(len(x)-1)[:,None] + np.arange(2)

x[ind]

array([[1, 2],
       [2, 3],
       [3, 4],
       [4, 5],
       [5, 6]])

Array into multiple arrays

6 Answers6

Linked