numpy.ndarray.strides — NumPy v2.2 Manual (original) (raw)

attribute

ndarray.strides#

Tuple of bytes to step in each dimension when traversing an array.

The byte offset of element (i[0], i[1], ..., i[n]) in an array _a_is:

offset = sum(np.array(i) * a.strides)

A more detailed explanation of strides can be found inThe N-dimensional array (ndarray).

Warning

Setting arr.strides is discouraged and may be deprecated in the future. numpy.lib.stride_tricks.as_strided should be preferred to create a new view of the same data in a safer way.

Notes

Imagine an array of 32-bit integers (each 4 bytes):

x = np.array([[0, 1, 2, 3, 4], [5, 6, 7, 8, 9]], dtype=np.int32)

This array is stored in memory as 40 bytes, one after the other (known as a contiguous block of memory). The strides of an array tell us how many bytes we have to skip in memory to move to the next position along a certain axis. For example, we have to skip 4 bytes (1 value) to move to the next column, but 20 bytes (5 values) to get to the same position in the next row. As such, the strides for the array x will be(20, 4).

Examples

import numpy as np y = np.reshape(np.arange(234), (2,3,4)) y array([[[ 0, 1, 2, 3], [ 4, 5, 6, 7], [ 8, 9, 10, 11]], [[12, 13, 14, 15], [16, 17, 18, 19], [20, 21, 22, 23]]]) y.strides (48, 16, 4) y[1,1,1] 17 offset=sum(y.strides * np.array((1,1,1))) offset/y.itemsize 17

x = np.reshape(np.arange(567*8), (5,6,7,8)).transpose(2,3,1,0) x.strides (32, 4, 224, 1344) i = np.array([3,5,2,2]) offset = sum(i * x.strides) x[3,5,2,2] 813 offset / x.itemsize 813