'and' (boolean) vs '&' (bitwise) - Why difference in behavior with lists vs numpy arrays?

Question

What explains the difference in behavior of boolean and bitwise operations on lists vs NumPy arrays?

I\'m confused about the appropriate use of

Answer 1

Operations with a Python list operate on the list. list1 and list2 will check if list1 is empty, and return list1 if it is, and list2 if it isn't. list1 + list2 will append list2 to list1, so you get a new list with len(list1) + len(list2) elements.

Operators that only make sense when applied element-wise, such as &, raise a TypeError, as element-wise operations aren't supported without looping through the elements.

Numpy arrays support element-wise operations. array1 & array2 will calculate the bitwise or for each corresponding element in array1 and array2. array1 + array2 will calculate the sum for each corresponding element in array1 and array2.

This does not work for and and or.

array1 and array2 is essentially a short-hand for the following code:

if bool(array1):
    return array2
else:
    return array1

For this you need a good definition of bool(array1). For global operations like used on Python lists, the definition is that bool(list) == True if list is not empty, and False if it is empty. For numpy's element-wise operations, there is some disambiguity whether to check if any element evaluates to True, or all elements evaluate to True. Because both are arguably correct, numpy doesn't guess and raises a ValueError when bool() is (indirectly) called on an array.