我更喜欢下一个提示: 将矩阵命名为matrix_a并使用column_number，例如:

import numpy as np
matrix_a = np.array([[1,2,3,4],[5,6,7,8],[9,10,11,12]])
column_number=2

# you can get the row from transposed matrix - it will be a column:
col=matrix_a.transpose()[column_number]

>>> import numpy as np
>>> A = np.array([[1,2,3,4],[5,6,7,8]])

>>> A
array([[1, 2, 3, 4],
    [5, 6, 7, 8]])

>>> A[:,2] # returns the third columm
array([3, 7])

参见:"numpy。“Arange”和“重塑”来分配内存

示例:(用矩阵(3x4)的形状分配数组)

nrows = 3
ncols = 4
my_array = numpy.arange(nrows*ncols, dtype='double')
my_array = my_array.reshape(nrows, ncols)

所有列从一个矩阵到一个新的列表:

N = len(matrix) 
column_list = [ [matrix[row][column] for row in range(N)] for column in range(N) ]

def get_col(arr, col):
    return map(lambda x : x[col], arr)

a = [[1,2,3,4], [5,6,7,8], [9,10,11,12],[13,14,15,16]]

print get_col(a, 3)

Python中的map函数是另一种方法。

[matrix[i][column] for i in range(len(matrix))]

>>> x = arange(20).reshape(4,5)
>>> x array([[ 0,  1,  2,  3,  4],
        [ 5,  6,  7,  8,  9],
        [10, 11, 12, 13, 14],
        [15, 16, 17, 18, 19]])

如果你想要第二列，你可以用

>>> x[:, 1]
array([ 1,  6, 11, 16])