修改numpy数组中的值

Question

Let's suppose a and b are given as below:-

a = np.arange(5)
b = [0,1,2]

What I want is that for indices excluding those in b ,values in a should be equal to -1. So in the above case a will be equal to

a = array([0, 1, 2, -1, -1])

There is a method which I am aware of ie

a[list(set(a)-set(b))] = -1

but takes too much time, and leads to too much complexity when actually writing the code. As always I am looking out for better methods than the above one. Feel free to use any tools required. Another example (just in case):- if

a = np.arange(12)
b = [3,5,6]

Then what I really want is a = array([-1, -1, -1, 3, -1, 5, 6, -1, -1, -1, -1, -1]) PS Don't worry a will always be of the form np.arange(int) and no value of b exceeds the length of a

Answer 1

>>> a = np.arange(12)
>>> b = [3,5,6]
>>> a[~np.in1d(np.arange(len(a)), b)] = -1
>>> a
array([-1, -1, -1,  3, -1,  5,  6, -1, -1, -1, -1, -1])

Answer 2

Well if the object is to create a range of values that are either the same as their index or -1 then it might be simpler to start with all -1 and add the data you want rather than the other way around.

>>> a = np.full(12, -1, dtype=int)
>>> b = [3, 5, 6]
>>> a[b] = b
>>> a
array([-1, -1, -1,  3, -1,  5,  6, -1, -1, -1, -1, -1])