[Numpy-discussion] views or copy

[josef.pktd at gmail.com]

I ran again into a problem where numpy created a view (which I didn't
realize) and an operation works differently on the view than if it
were a copy.

I try to construct an example array, which, however, is only a view

>>> x,y = np.mgrid[0:3,0:3]
>>> xx = np.vstack((x.flatten(), y.flatten(), np.ones(9))).T
>>> xx
array([[ 0.,  0.,  1.],
       [ 0.,  1.,  1.],
       [ 0.,  2.,  1.],
       [ 1.,  0.,  1.],
       [ 1.,  1.,  1.],
       [ 1.,  2.,  1.],
       [ 2.,  0.,  1.],
       [ 2.,  1.,  1.],
       [ 2.,  2.,  1.]])
>>> xx.flags
  C_CONTIGUOUS : False
  F_CONTIGUOUS : True
  OWNDATA : False
  WRITEABLE : True
  ALIGNED : True
  UPDATEIFCOPY : False
>>> xx.base
array([[ 0.,  0.,  0.,  1.,  1.,  1.,  2.,  2.,  2.],
       [ 0.,  1.,  2.,  0.,  1.,  2.,  0.,  1.,  2.],
       [ 1.,  1.,  1.,  1.,  1.,  1.,  1.,  1.,  1.]])
>>> xx == xx.base
False

When I convert it to a view as a structured array, it produces a
strange result. I didn't get what I thought I should get

>>> xx.view([('',xx.dtype)]*xx.shape[1])
array([[(0.0, 0.0, 0.0), (0.0, 1.0, 2.0), (1.0, 1.0, 1.0)],
       [(1.0, 1.0, 1.0), (0.0, 1.0, 2.0), (1.0, 1.0, 1.0)],
       [(2.0, 2.0, 2.0), (0.0, 1.0, 2.0), (1.0, 1.0, 1.0)]],
      dtype=[('f0', '<f8'), ('f1', '<f8'), ('f2', '<f8')])

if I make a copy and then construct a view as structured array, I get
what I want

>>> xx2 = xx.copy()
>>> xx2.view([('',xx2.dtype)]*xx2.shape[1])
array([[(0.0, 0.0, 1.0)],
       [(0.0, 1.0, 1.0)],
       [(0.0, 2.0, 1.0)],
       [(1.0, 0.0, 1.0)],
       [(1.0, 1.0, 1.0)],
       [(1.0, 2.0, 1.0)],
       [(2.0, 0.0, 1.0)],
       [(2.0, 1.0, 1.0)],
       [(2.0, 2.0, 1.0)]],
      dtype=[('f0', '<f8'), ('f1', '<f8'), ('f2', '<f8')])

Are there rules for this behavior or a description in the docs,
because my mistakes in this are quite difficult to debug? Or do I have
to make a copy by default as in matlab?

Josef