Inconsistent behavior if fill data outside of array boundaries

Sorry, if it is duplicate, but I didn’t succeed to formulate search query properly.

When inserting data out of array boundaries, sometimes numpy don’t throw an error

Reproducing code example:

import numpy as np
a =  np.array([1,2,3])
a[10:12] = [10, 10]   # raises a ValueError
a[10:11] = [10]   # does nothing, no exception, just silently passes

I expect both lines behave consistently, f.ex. throw the same ValueError exception

Numpy/Python version information:

numpy 1.18.1 Python 3.6.4 (v3.6.4:d48ecebad5, Dec 18 2017, 21:07:28) [GCC 4.2.1 (Apple Inc. build 5666) (dot 3)]

[eric-wieser]

You’re right, that page could do with an example with a length-0 dimension.

The key to understanding is that when “expanding” from 1 to N, 0 is a legal value for N just like any other value.

[eric-wieser]

Doesn’t my comment above cover that? In a situation where the code is “broadcast this value to all elements satisfying this condition”, you want the broadcast to succeed even if sometimes no elements satisfy the condition