Shift elements in a numpy array

Question

Following-up from this question years ago, is there a canonical \"shift\" function in numpy? I don\'t see anything from the documentation.

Here\'s a simple version o

Answer 1

Not numpy but scipy provides exactly the shift functionality you want,

import numpy as np
from scipy.ndimage.interpolation import shift

xs = np.array([ 0.,  1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9.])

shift(xs, 3, cval=np.NaN)

where default is to bring in a constant value from outside the array with value cval, set here to nan. This gives the desired output,

array([ nan, nan, nan, 0., 1., 2., 3., 4., 5., 6.])

and the negative shift works similarly,

shift(xs, -3, cval=np.NaN)

Provides output

array([  3.,   4.,   5.,   6.,   7.,   8.,   9.,  nan,  nan,  nan])

Answer 2

You can also do this with Pandas:

Using a 2356-long array:

import numpy as np

xs = np.array([...])

Using scipy:

from scipy.ndimage.interpolation import shift

%timeit shift(xs, 1, cval=np.nan)
# 956 µs ± 77.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

Using Pandas:

import pandas as pd

%timeit pd.Series(xs).shift(1).values
# 377 µs ± 9.42 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

In this example, using Pandas was about ~8 times faster than Scipy

Answer 3

There is no single function that does what you want. Your definition of shift is slightly different than what most people are doing. The ways to shift an array are more commonly looped:

>>>xs=np.array([1,2,3,4,5])
>>>shift(xs,3)
array([3,4,5,1,2])

However, you can do what you want with two functions.
Consider a=np.array([ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9.]):

def shift2(arr,num):
    arr=np.roll(arr,num)
    if num<0:
         np.put(arr,range(len(arr)+num,len(arr)),np.nan)
    elif num > 0:
         np.put(arr,range(num),np.nan)
    return arr
>>>shift2(a,3)
[ nan  nan  nan   0.   1.   2.   3.   4.   5.   6.]
>>>shift2(a,-3)
[  3.   4.   5.   6.   7.   8.   9.  nan  nan  nan]

After running cProfile on your given function and the above code you provided, I found that the code you provided makes 42 function calls while shift2 made 14 calls when arr is positive and 16 when it is negative. I will be experimenting with timing to see how each performs with real data.

Answer 4

Benchmarks & introducing Numba

1. Summary

• The accepted answer (scipy.ndimage.interpolation.shift) is the slowest solution listed in this page.
• Numba (@numba.njit) gives some performance boost when array size smaller than ~25.000
• "Any method" equally good when array size large (>250.000).
• The fastest option really depends on
      (1)  Length of your arrays
      (2)  Amount of shift you need to do.
• Below is the picture of the timings of all different methods listed on this page (2020-07-11), using constant shift = 10. As one can see, with small array sizes some methods are use more than +2000% time than the best method.

2. Detailed benchmarks with the best options

• Choose shift4_numba (defined below) if you want good all-arounder

3. Code

3.1 shift4_numba

• Good all-arounder; max 20% wrt. to the best method with any array size
• Best method with medium array sizes: ~ 500 < N < 20.000.
• Caveat: Numba jit (just in time compiler) will give performance boost only if you are calling the decorated function more than once. The first call takes usually 3-4 times longer than the subsequent calls.

import numba

@numba.njit
def shift4_numba(arr, num, fill_value=np.nan):
    if num >= 0:
        return np.concatenate((np.full(num, fill_value), arr[:-num]))
    else:
        return np.concatenate((arr[-num:], np.full(-num, fill_value)))

3.2. shift5_numba

• Best option with small (N <= 300.. 1500) array sizes. Treshold depends on needed amount of shift.
• Good performance on any array size; max + 50% compared to the fastest solution.
• Caveat: Numba jit (just in time compiler) will give performance boost only if you are calling the decorated function more than once. The first call takes usually 3-4 times longer than the subsequent calls.

import numba

@numba.njit
def shift5_numba(arr, num, fill_value=np.nan):
    result = np.empty_like(arr)
    if num > 0:
        result[:num] = fill_value
        result[num:] = arr[:-num]
    elif num < 0:
        result[num:] = fill_value
        result[:num] = arr[-num:]
    else:
        result[:] = arr
    return result

3.3. shift5

• Best method with array sizes ~ 20.000 < N < 250.000
• Same as shift5_numba, just remove the @numba.njit decorator.

4 Appendix

4.1 Details about used methods

• shift_scipy: scipy.ndimage.interpolation.shift (scipy 1.4.1) - The option from accepted answer, which is clearly the slowest alternative.
• shift1: np.roll and out[:num] xnp.nan by IronManMark20 & gzc
• shift2: np.roll and np.put by IronManMark20
• shift3: np.pad and slice by gzc
• shift4: np.concatenate and np.full by chrisaycock
• shift5: using two times result[slice] = x by chrisaycock
• shift#_numba: @numba.njit decorated versions of the previous.

The shift2 and shift3 contained functions that were not supported by the current numba (0.50.1).

4.2 Other test results

4.2.1 Relative timings, all methods

• Relative timings, 10% shift, all methods
• Relative timings, constant shift (10), all methods

4.2.2 Raw timings, all methods

• Raw timings, constant shift (10), all methods
• Raw timings, 10% shift, all methods

4.2.3 Raw timings, few best methods

• Raw timings with small arrays, constant shift (10), few best methods
• Raw timings with small arrays, 10% shift, few best methods
• Raw timings with large arrays, constant shift (10), few best methods
• Raw timings with large arrays, 10% shift, few best methods

Answer 5

If you want a one-liner from numpy and aren't too concerned about performance, try:

np.sum(np.diag(the_array,1),0)[:-1]

Explanation: np.diag(the_array,1) creates a matrix with your array one-off the diagonal, np.sum(...,0) sums the matrix column-wise, and ...[:-1] takes the elements that would correspond to the size of the original array. Playing around with the 1 and :-1 as parameters can give you shifts in different directions.

Answer 6

One way to do it without spilt the code into cases

with array:

def shift(arr, dx, default_value):
    result = np.empty_like(arr)
    get_neg_or_none = lambda s: s if s < 0 else None
    get_pos_or_none = lambda s: s if s > 0 else None
    result[get_neg_or_none(dx): get_pos_or_none(dx)] = default_value
    result[get_pos_or_none(dx): get_neg_or_none(dx)] = arr[get_pos_or_none(-dx): get_neg_or_none(-dx)]     
    return result

with matrix it can be done like this:

def shift(image, dx, dy, default_value):
    res = np.full_like(image, default_value)

    get_neg_or_none = lambda s: s if s < 0 else None
    get_pos_or_none = lambda s : s if s > 0 else None

    res[get_pos_or_none(-dy): get_neg_or_none(-dy), get_pos_or_none(-dx): get_neg_or_none(-dx)] = \
        image[get_pos_or_none(dy): get_neg_or_none(dy), get_pos_or_none(dx): get_neg_or_none(dx)]
    return res