Question

我有一个numpy数组，其形状（n，）是一个包含多个元素的元组数组。有没有办法让我快速添加一个列到其中一个元组和＆amp;它仍然具有相同的形状（n，）？

Answer 1

由于这个问题没有答案，我写了一个可能有帮助的片段。您可以使用np.zeros或np.hstack向阵列添加新列。

import numpy as np

def main():
    my_tuples = np.array(((1,-2,3),(4,5,6))) 
    element = (7,8) #suppose you want to add this to my_tuples column

    #Solution 1
    print('shape before adding column', my_tuples.shape)
    new_tuples = np.zeros((my_tuples.shape[0],my_tuples.shape[1]+1));
    new_tuples[:,:-1] = my_tuples
    new_tuples[:,-1] = element
    print('column to be added ', element)
    print('shape after adding column using np.zeros', new_tuples.shape)
    print(new_tuples)

    #Solution 2
    new_tuples = np.hstack((my_tuples,np.zeros((my_tuples.shape[0],1))))
    new_tuples[:,-1] = element
    print('shape after adding column using np.hstack', new_tuples.shape)
    print(new_tuples)


if __name__ == '__main__':
    main()

<强>输出：

shape before adding column (2, 3)
column to be added  (7, 8)
shape after adding column using np.zeros (2, 4)
[[ 1. -2.  3.  7.]
 [ 4.  5.  6.  8.]]
shape after adding column using np.hstack (2, 4)
[[ 1. -2.  3.  7.]
 [ 4.  5.  6.  8.]]

Answer 2

我认为我找到了一个解决方案，它为结构数组添加了一个列：

 from numpy.lib import recfunctions
 a = recfunctions.append_fields(old_data,'new_column', 
                                new_array,dtypes=np.float64,
                                usemask=False)

Pandas Numpy添加了一个专栏

2 个答案: