Named dtype array: Difference between a[0]['name'] and a['name'][0]? (Trac #2068)

[numpy-gitbot]

Original ticket http://projects.scipy.org/numpy/ticket/2068 on 2012-02-27 by trac user nschloe, assigned to unknown.

Hi!

I came across the following oddity in numpy which may or may not be a bug:

import numpy as np
dt = np.dtype([('tuple', (int, 2))])
a = np.zeros(3, dt)
type(a['tuple'][0])  % ndarray
type(a[0]['tuple'])  % ndarray

a['tuple'][0] = (1,2)  % ok
a[0]['tuple'] = (1,2)  % ValueError: shape-mismatch on array construction

I would have expected that both of the options below work.
Opinions?

[numpy-gitbot]

@teoliphant wrote on 2012-02-27

This is a known quirk. a[0] is a numpy.void array-scalar type (basically something like a named tuple). It would take some work for this to be possible and is not a priority at this point. One possibility would be if a[0] instead returned a 0-d array whenever the dtype was not simple.

[ahaldane]

This was fixed by one of my PRs. Closing.

[ahaldane]

(in #5947)