SO I have two arrays
a=np.array([1,2,3])
b=np.array(['a','b','c'])
That I need to combine such that I get
array([1,'a'],[2,'b'],[3,'c'])
I thought a simple
np.stack((a,b),axis=1)
would do it, but it turns everything into a string. How do I avoid that?
array([['1', 'a'],
['2', 'b'],
['3', 'c']],
dtype='<U21')
You can put them in a structured array
In [104]: a=np.array([1,2,3])
...: b=np.array(['a','b','c'])
In [105]: arr = np.empty(a.shape[0], dtype='int,U4')
In [106]: arr
Out[106]:
array([(0, ''), (0, ''), (0, '')],
dtype=[('f0', '<i4'), ('f1', '<U4')])
In [107]: arr['f0']=a
In [108]: arr['f1']=b
In [109]: arr
Out[109]:
array([(1, 'a'), (2, 'b'), (3, 'c')],
dtype=[('f0', '<i4'), ('f1', '<U4')])
In [111]: arr['f1']
Out[111]:
array(['a', 'b', 'c'],
dtype='<U4')
https://docs.scipy.org/doc/numpy/user/basics.rec.html
If that isn't useful, stick with separate arrays.
You can't avoid it, since they need to be of the same type. However, you can do something like this:
import numpy as np
a = np.array([1,2,3])
b = np.array(['a','b','c'])
x = np.stack((a,b),axis=1)
x[0][0].astype(np.int)
>> 1
x[0][1]
>> 'a'
But it's not the nicest code.
numpyarray have to be of same type.