Why numpy data type kind returns void when it was created as float64?

Question

I have this code:

>>> d = np.dtype([("pos", np.float64, 3)])
>>> d[0].kind
'V'

Why does it return 'V' instead of 'f'? In the full code, I need to know if the field corresponds to an integer, float, string...

Answer 1

d[0].kind does not return 'f' because it is not a floating-point dtype: it is a structured dtype with a floating point base.

There are several other attributes of structured dtypes that you might be able to inspect depending on your particular goal. For example:

>>> d[0].base
dtype('float64')
>>> d[0].subdtype[0].kind
'f'