If I use np.concatenate to join this 1x5 array
array([True , False, False, True , True ])
with this 1x5 array
array([4.753, 1.202, 2.296, 1.668, 3.35 ])
The booleans are changed to integers:
array([[1. , 0. , 0. , 1. , 1. ],
[4.753, 1.202, 2.296, 1.668, 3.35 ]])
Why? how can I concatenate them without affecting the booleans?
Ideally, the output can be a pandas DataFrame
Something like this?
In [445]: import numpy as np
...: import pandas as pd
In [446]: data = {'Booleans': np.array([True, False, False, True, True]),
...: 'Floats': np.array([4.753, 1.202, 2.296, 1.668, 3.35])}
In [447]: df = pd.DataFrame(data)
In [448]: df
Out[448]:
Booleans Floats
0 True 4.753
1 False 1.202
2 False 2.296
3 True 1.668
4 True 3.350
You may find this link useful: Different ways to create Pandas Dataframe.
One way is to set dtype of the individual arrays as object (you just need one of them to be object type to result in an object type array)
a = np.array([[True , False, False, True , True]], dtype=object)
b = np.array([[4.753, 1.202, 2.296, 1.668, 3.35]])
np.concatenate([a,b])
array([[True, False, False, True, True],
[4.753, 1.202, 2.296, 1.668, 3.35]], dtype=object)
You can use arr.astype() to convert each array to object type after defining them as well.
a.astype(object)
#array([[True, False, False, True, True]], dtype=object)
