Subclassing the numpy array, where all form change operations return a normal array

Question

I have a subclass of an array where some additional attributes are only valid for the original form of the object. Is there a way to make sure that all array shape changing operations return a normal numpy array instead of an instance of my class?

I already wrote array_wrap, but that doesn't seem to affect operations like np.mean , np.sum or np.rollaxis . All this just returns an instance of my class.

 import numpy as np class NewArrayClass(np.ndarray): __array_priority__ = 3.0 def __array_wrap__(self, out_arr, context=None): if out_arr.shape == self.shape: out = out_arr.view(new_array) # Do a bunch of class dependant initialization and attribute copying. # ... return out else: return np.asarray(out_arr) A = np.arange(10) A.shape = (5, 2) A = arr.view(NewArrayClass) # Would like this to be np.ndarray, but get new_array_class. print type(np.sum(A, 0)) 

I suppose I need to do something in __new__ or __array_finalize__ , but I don't know what.

Update: After carefully reading the numpy documentation for the subclass ( http://docs.scipy.org/doc/numpy/user/basics.subclassing.html ), all operations of changing the shape of the array perform the operation “new from the template”. So the question is, how to make the "new from template" operation return ndarray instances instead of instances of my class. As far as I can tell, __new__ never called inside these functions.

Alternative: Assuming this is not possible, how do I at least identify in __array_finalize__ new template operation (as opposed to viewing)? This, at least, would allow me to dereference some attributes that are copied by reference. I could also set a flag or something telling the new instance that its form is invalid.

Answer 1

If you do not introduce new members into your NewArrayClass instances, you can reassign the __class__ attribute of the returned instances.

 A.__class__ = np.ndarray 

That is why you would like to do that. Do you have stringent type checks elsewhere? You would go much more with a duck set.