I created a NumPy array,
a = numpy.array([[1,2,3][4,5,6]])
I want to make the array look like this [[1,4],[2,5],[3,6]] and also after I make change I want to return to the original structure.
Is there a NumPy command to run a function on all values, like a[0] * 2?
The result should be
[[2,8][2,5][3,6]
You want to transpose the array (think matrices). Numpy arrays have a method for that:
a = np.array([[1,2,3],[4,5,6]])
b = a.T # or a.transpose()
But note that b is now a view of a; if you change b, a changes as well (this saves memory and time otherwise spent copying).
You can change the first column of b with
b[0] *= 2
Which gives you the result you want, but a has also changed! If you don't want that, use
b = a.T.copy()
If you do want to change a, note that you can also immediately change the values you want in a itself:
a[:, 0] *= 2
You can use zip on the ndarray and pass it to numpy.array:
In [36]: a = np.array([[1,2,3], [4,5,6]])
In [37]: b = np.array(zip(*a))
In [38]: b
Out[38]:
array([[1, 4],
[2, 5],
[3, 6]])
In [39]: b*2
Out[39]:
array([[ 2, 8],
[ 4, 10],
[ 6, 12]])
Use numpy.column_stack for a pure NumPy solution:
In [44]: b = np.column_stack(a)
In [45]: b
Out[45]:
array([[1, 4],
[2, 5],
[3, 6]])
In [46]: b*2
Out[46]:
array([[ 2, 8],
[ 4, 10],
[ 6, 12]])
