Error when converting int to numpy array?

Question

I encounter the following errors when converting an integer to a numpy array:

a = 150
b = np.array(a)
b[0]
*** IndexError: too many indices for array: array is 0-dimensional, but 1 were indexed

I expect the output of b[0] is 150, but get the IndexError. Why?

the shape of your array is (). you can use np.array([a]) if you want to define a one dimentional array. — Odeaxcsh
– Odeaxcsh, Commented Aug 18, 2022 at 10:37
Yes. And I don't understand the output of b is array(150) instead of array([150]). What does it mean? — f. c.
– f. c., Commented Aug 18, 2022 at 10:41
notice that the output is based on the input shape. when you pass an scalar like 150, it's reasonable for output that have a dimensional of 0. by the way you can use indexing like b[()]. — Odeaxcsh
– Odeaxcsh, Commented Aug 18, 2022 at 12:34
b[()] indexes the 0d array with a 0 element tuple. np.array([150])[0] indexes with a 1d array with 1 index. b.item() is another way of getting the element (from either). — hpaulj
– hpaulj, Commented Aug 18, 2022 at 15:26

doctorlove · Accepted Answer · 2022-08-18 10:41:50Z

2

You have in effect created a scalar or zero dimensional array.

If you need to index into this, try

b = np.array([a])

answered Aug 18, 2022 at 10:41

doctorlove

19.4k3 gold badges49 silver badges65 bronze badges

Sign up to request clarification or add additional context in comments.

1 Answer 1