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
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)
On Mon, Oct 17, 2011 at 6:17 AM, Pauli Virtanen <pav@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
17.10.2011 15:48, josef.pktd@gmail.com kirjoitti:
On Mon, Oct 17, 2011 at 6:17 AM, Pauli Virtanen <pav@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
On Mon, Oct 17, 2011 at 10:18 AM, Pauli Virtanen <pav@iki.fi> wrote:
17.10.2011 15:48, josef.pktd@gmail.com kirjoitti:
On Mon, Oct 17, 2011 at 6:17 AM, Pauli Virtanen <pav@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
participants (3)
-
Alex van der Spek -
josef.pktd＠gmail.com -
Pauli Virtanen