有人知道如何在Python中从多维数组中提取列吗?


当前回答

只要使用转置(),就可以像求行一样简单地求列

matrix=np.array(originalMatrix).transpose()
print matrix[NumberOfColumns]

其他回答

如果你在Python中有一个二维数组(不是numpy),你可以像这样提取所有的列,

data = [
['a', 1, 2], 
['b', 3, 4], 
['c', 5, 6]
]

columns = list(zip(*data))

print("column[0] = {}".format(columns[0]))
print("column[1] = {}".format(columns[1]))
print("column[2] = {}".format(columns[2]))

执行这段代码会得到,

>>> print("column[0] = {}".format(columns[0]))
column[0] = ('a', 'b', 'c')

>>> print("column[1] = {}".format(columns[1]))
column[1] = (1, 3, 5)

>>> print("column[2] = {}".format(columns[2]))
column[2] = (2, 4, 6)

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]

假设我们有nxm矩阵(n行m列)5行4列

matrix = [[1,2,3,4],[5,6,7,8],[9,10,11,12],[13,14,15,16],[17,18,19,20]]

要在python中提取列,我们可以像这样使用列表推导式

[ [row[i] for row in matrix] for in range(4) ]

你可以用矩阵的列数来替换4。 结果是

,10,14,18,5,9,13,17 [[1], [2], [3,7,11,15,19], [4,8,12,16,20]]

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

col1 = [val[1] for val in array]
col2 = [val[2] for val in array]
col3 = [val[3] for val in array]
col4 = [val[4] for val in array]
print(col1)
print(col2)
print(col3)
print(col4)

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

点击这里查看详情!

a = [[1, 2], [2, 3], [3, 4]]
a2 = zip(*a)
a2[0]

它和上面的是一样的,只是它更整洁一些 zip可以完成这项工作,但需要单个数组作为参数,*a语法将多维数组解压缩为单个数组参数