You could vectorize the function:
>>> import numpy
>>>
>>> class Obj(object):
... def __init__(self, x, y):
... self.x = x
... self.y = y
...
>>> arr = numpy.array([Obj(1, 2), Obj(3, 4), Obj(5, 6)])
>>>
>>> vectorized_x = numpy.vectorize(lambda obj: obj.x)
>>>
>>> vectorized_x(arr)
array([1, 3, 5])
Although I'm not sure if you should really store a NumPy array of Python objects in the first place. Vectorize is no more efficient than a Python loop. It would be more efficient to store an (n+1)-D array, as we could easily extract the contents simply by slicing which is a native operation, e.g.
>>> a = numpy.array([[(1, 2), (3, 4), (5, 6)], [(7, 8), (9, 10), (11, 12)], [(-13, -14), (-15, -16), (-17, -18)]])
>>> a[:,:,0]
array([[ 1, 3, 5],
[ 7, 9, 11],
[-13, -15, -17]])
>>> a[:,:,1]
array([[ 2, 4, 6],
[ 8, 10, 12],
[-14, -16, -18]])
