NumPy | roll method
Start your free 7-days trial now!
Numpy's roll(~)
method shifts an array along the specified axis.
Parameters
1. a
| array_like
✜ The array to perform the method on.
2. shift
link | integer
or tuple
of integers
✜ The desired number of shifts.
3. axis
link | number
| optional
✜ The axis along which to shift the input array.
Return value
A Numpy array with elements shifted by the specified amount.
Examples
Rolling a flat array
To shift a flat array by 1, use the following:
np.roll([1,2,3,4], 1)
array([4, 1, 2, 3])
Notice how the final value of the array (i.e. 4 in this case), got placed in the very front.
Rolling a 2D array
By value
To shift a 2D array by 1, use the following:
np.roll([[1,2], [3,4]], 1)
array([[4, 1], [2, 3]])
Essentially, what we are doing is as follows:
By row
To shift a 2D array by 1 row-wise, add the axis=0
parameter:
np.roll([[1,2], [3,4], [5,6]], 1, axis=0)
array([[5, 6], [1, 2], [3, 4]])
Essentially, we're doing the following:
By column
To shift a 2D array by 1 column-wise, add the axis=1
parameter:
np.roll([[1,2,3], [4,5,6]], 1, axis=1)
array([[3, 1, 2], [6, 4, 5]])
Essentially, we're doing the following: