[Numpy-discussion] dtyping with .astype()
Beginner's question?
I have this dictionary dtypes of names and types:
dtypes
{'names': ['col1', 'col2', 'col3', 'col4', 'col5'], 'formats': [type
'numpy.float16', type 'numpy.float16', type 'numpy.float16', type
'numpy.float16', type 'numpy.float16']}
and this array y
y
array([[ 0, 1, 2, 3, 4],
[ 5, 6, 7, 8, 9],
[10, 11, 12, 13, 14],
[15, 16, 17, 18, 19],
[20, 21, 22, 23, 24],
[25, 26, 27, 28, 29],
[30, 31, 32, 33, 34],
[35, 36, 37, 38, 39],
[40, 41, 42, 43, 44],
[45, 46, 47, 48, 49]])
But:
z=y.astype(dtypes)
gives me a confusing result. I only asked to name the columns and change their
types to half precision floats.
What am I missing? How to do this?
Thank you in advance,
Alex van der Spek
___
NumPy-Discussion mailing list
NumPy-Discussion@scipy.org
http://mail.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] dtyping with .astype()
13.10.2011 12:59, Alex van der Spek kirjoitti: gives me a confusing result. I only asked to name the columns and change their types to half precision floats. Structured arrays shouldn't be thought as an array with named columns, as they are somewhat different. What am I missing? How to do this? np.rec.fromarrays(arr.T, dtype=dt) ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] dtyping with .astype()
On Mon, Oct 17, 2011 at 6:17 AM, Pauli Virtanen p...@iki.fi wrote: 13.10.2011 12:59, Alex van der Spek kirjoitti: gives me a confusing result. I only asked to name the columns and change their types to half precision floats. Structured arrays shouldn't be thought as an array with named columns, as they are somewhat different. What am I missing? How to do this? np.rec.fromarrays(arr.T, dtype=dt) y.astype(float16).view(dt) ? Josef ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] dtyping with .astype()
17.10.2011 15:48, josef.p...@gmail.com kirjoitti: On Mon, Oct 17, 2011 at 6:17 AM, Pauli Virtanen p...@iki.fi wrote: [clip] What am I missing? How to do this? np.rec.fromarrays(arr.T, dtype=dt) y.astype(float16).view(dt) I think this will give surprises if the original array is not in C-order. Pauli ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] dtyping with .astype()
On Mon, Oct 17, 2011 at 10:18 AM, Pauli Virtanen p...@iki.fi wrote: 17.10.2011 15:48, josef.p...@gmail.com kirjoitti: On Mon, Oct 17, 2011 at 6:17 AM, Pauli Virtanen p...@iki.fi wrote: [clip] What am I missing? How to do this? np.rec.fromarrays(arr.T, dtype=dt) y.astype(float16).view(dt) I think this will give surprises if the original array is not in C-order. I forgot about those, dangerous if the array is square, otherwise it raises an error if it is in F-order maybe np.asarray(y, np.float16, order='C').view(dt) if I don't like record arrays Josef Pauli ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
