Numpy 数组复合排序——mX4列，根据多列组合排序

Numpy 数组复合排序——mX4列，根据多列组合排序

一、 Numpy 仅根据某列对 array（mX4）整体排序用argsort

import numpy as np

data = np.array([[1, 5, 2, 4, 10], [9, 6, 8, 4, 5], [2, 3, 4, 6, 3], [7, 6, 8, 4, 6]])
print(' before---- ')
print(data)
data = data[data[:, 0].argsort()]
print('根据第1列排序 after: ')
print(data)

运行结果：

二、mX4根据多列组合排序（先根据x排序，x相同根据y排序，y相同根据z排序）

分以下三步进行，源代码见底部：

1. np array(array) 转array(tuple) [(1,2,3,4),(5,6,7,8),…] ----> [(1,2,3,4),(5,6,7,8),…]

m2A = []
    for i in range(0, len(inFile)):
        m2A.append((inFile[i][0], inFile[i][1], inFile[i][2], inFile[i][3]))

1. np 对array(tuple) 进行 多列组合排序（先根据x排序，再根据y排序，再根据z排序） m2A: [(1,2,3,4),(5,6,7,8),…]

dtype = [('x', float), ('y', float), ('z', float), ('i', int)]
tuple1 = np.array(m2A, dtype)
tuple1 = np.sort(tuple1, order=['x', 'y', 'z'])

1. np array(tuple) 转 array(array)

   该过程包括俩步：
   (1) np array 递归转为 tuple [(1,2,3,4),(5,6,7,8),…] ----> ((1,2,3,4),(5,6,7,8),…)
   (2) np tuple 转 array [(1,2,3,4),(5,6,7,8),…] —> [[1,2,3,4],[5,6,7,8],…]

inFile = np.array(totuple(tuple1))   # 先整体转为元组，再转多维数组

import numpy as np

np.set_printoptions(suppress=True)

def totuple(a):
    try:
        return tuple(totuple(i) for i in a)
    except TypeError:
        return a

data = np.array([[423999.72382051, 4421127.27096031, 83.313, 7.],
                  [423993.12782051, 4421123.49396031, 83.276, 23.],
                  [423993.05882051, 4421122.56596031, 84.28, 82.],
                  [423992.95382051, 4421123.20896031, 83.293, 15.],
                  [423993.08082051, 4421122.47896031, 84.429, 100.]])

print('排序前: ',data.shape,data)

m2A = []
for i in range(0, len(data)):
    m2A.append((data[i][0], data[i][1], data[i][2], data[i][3]))

dtype = [('x', float), ('y', float), ('z', float), ('i', int)]
tuple1 = np.array(m2A, dtype)
tuple1 = np.sort(tuple1, order=['x', 'y', 'z'])
print(tuple1.shape)

inFile = np.array(totuple(tuple1))   # 转为元组，转array

print('排序后：', data)