有人知道如何在Python中从多维数组中提取列吗?
当前回答
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]
其他回答
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]]
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) ]
>>> 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)
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)
如果你喜欢map-reduce风格的python, itemgetter操作符也会有帮助,而不是列表推导式,为了一点变化!
# tested in 2.4
from operator import itemgetter
def column(matrix,i):
f = itemgetter(i)
return map(f,matrix)
M = [range(x,x+5) for x in range(10)]
assert column(M,1) == range(1,11)