[Numpy-discussion] assignment to array.dtype

[Jan Strube]

Hi list,
maybe this is a really stupid idea, and I don't want to advertise this, but
what actually happens when I reassign the dtype of an object?
Does it return the same as array.view?

say I have the following code
In [64]: b
Out[64]:
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11],
       [12, 13, 14]])
<snip />
In [69]: t = N.dtype(zip(('a', 'b', 'c'), (N.int32, N.int32, N.int32)))

In [70]: b.view(t)
Out[70]:
array([[(0, 1, 2)],
       [(3, 4, 5)],
       [(6, 7, 8)],
       [(9, 10, 11)],
       [(12, 13, 14)]],
      dtype=[('a', '<i4'), ('b', '<i4'), ('c', '<i4')])
<snip />
In [75]: b.dtype=t

In [76]: b
Out[76]:
array([[(0, 1, 2)],
       [(3, 4, 5)],
       [(6, 7, 8)],
       [(9, 10, 11)],
       [(12, 13, 14)]],
      dtype=[('a', '<i4'), ('b', '<i4'), ('c', '<i4')])

So in this case is b.dtype = t the same as b = b.view(t)?
Or is this paradigm deprecated for some reason?
Just curious.
Thanks,
    Jan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20070322/f2bde6f2/attachment.html>