Wnma's Blogs
0%

Numpy的70个操作

发表于 更新于 分类于

这篇文章主要介绍了Numpy从入门到进阶的操作, 翻译文章，便于自己使用时查阅

原文链接

注释及输出前带有 #

1-10

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
import numpy as np

# 查看numpy的版本
np.__version__
# 1.13.3

# 创建一个一维数据
arr = np.arange(10)
arr
# array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

# 创建一个布尔型数组
# 方法一：
np.full((3, 3), True, dtype=bool)
# 方法二：
np.ones((3, 3), dtype=bool)
# array([[ True,  True,  True],
#        [ True,  True,  True],
#        [ True,  True,  True]], dtype=bool)

# 提取数组中为奇数的数字
arr = np.arange(10) # 原来的数组
# array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
arr[arr % 2 == 1]
# array([1, 3, 5, 7, 9])

# 替换数组中为奇数的数字为-1
arr = np.arange(10) # 原来的数组
# array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
# 方法一：
arr[arr % 2 == 1] = -1
# 方法二, 更加安全的操作方法（这里算新的一个操作）：
out = np.where(arr % 2 == 1, -1, arr)
out
# array([ 0, -1,  2, -1,  4, -1,  6, -1,  8, -1])

# 数组维度变换
arr = np.arange(10) # 原来的数组
arr.reshape(2, -1)
arr
# array([[0, 1, 2, 3, 4],
#        [5, 6, 7, 8, 9]])

# 纵向堆叠两个数组
a = np.arange(10).reshape(2, -1) # 原来的数组a
b = np.repeat(1, 10).reshape(2, -1) # 原来的数组b
# 方法一：
np.concatenate([a, b], axis=0)
# 方法二:
np.vstack([a, b])
# 方法三：
np.r_[a, b]
# 三种方法的输出
# array([[0, 1, 2, 3, 4],
#        [5, 6, 7, 8, 9],
#        [1, 1, 1, 1, 1],
#        [1, 1, 1, 1, 1]])

# 横向堆叠两个数组
a = np.arange(10).reshape(2, -1) # 原来的数组a
b = np.repeat(1, 10).reshape(2, -1) # 原来的数组b
# 方法一：
np.concatenate([a, b], axis=1)
# 方法二:
np.hstack([a, b])
# 方法三：
np.c_[a, b]
# 三种方法的输出
# array([[0, 1, 2, 3, 4, 1, 1, 1, 1, 1],
#        [5, 6, 7, 8, 9, 1, 1, 1, 1, 1]])

# 根据原有的数组，自由拓展
a = np.array([1, 2, 3]) # 原来的数组
np.r_[np.repeat(a, 3), np.tile(a, 3)]
# array([1, 1, 1, 2, 2, 2, 3, 3, 3, 1, 2, 3, 1, 2, 3, 1, 2, 3])

11-20

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
import numpy as np

# 获取两个数组中公共的元素
a = np.array([1, 2, 3, 2, 3, 4, 3, 4, 5, 6])  # 原来的数组a
b = np.array([7, 2, 10, 2, 7, 4, 9, 4, 9, 8]) # 原来的数组b
np.intersect1d(a, b)
# array([2, 4])

# 在a数组中移去a、b数组中公共的元素
a = np.array([1, 2, 3, 4, 5])	# 原来的数组a
b = np.array([5, 6, 7, 8, 9])	# 原来的数组b
np.setdiff1d(a, b)
# array([1, 2, 3, 4])

# 获取两个数组中公共元素（索引相同）的索引
a = np.array([1, 2, 3, 2, 3, 4, 3, 4, 5, 6])  # 原来的数组a
b = np.array([7, 2, 10, 2, 7, 4, 9, 4, 9, 8]) # 原来的数组b
np.where(a == b)
# (array([1, 3, 5, 7], dtype=int64),)

# 获取指定数值区间的元素
a = np.arange(15) # 原来的数组
# 方法一
index = np.where((a >= 5) & (a <= 10))
a[index]
# 方法二
index = np.where(np.logical_and(a>=5, a<=10))
a[index]
# 方法三
a[(a>= 5) & (a <= 10)]
# array([ 5,  6,  7,  8,  9, 10])

# numpy运用函数
def maxx(x, y):		# 需要使用的函数
    if x >= y:
        return x
    else:
        return y

a = np.array([5, 7, 9, 8, 6, 4, 5]) # 原来的数组a
b = np.array([6, 3, 4, 8, 9, 7, 1]) # 原来的数组b
pair_max = np.vectorize(maxx, dtypes=[float])
pair_max(a, b)
# array([ 6.,  7.,  9.,  8.,  9.,  7.,  5.])

# 二维数组交换两行
arr = np.arange(9).reshape(3,3)	# 原来的数组
arr[:, [1, 0, 2]]
# array([[1, 0, 2],
#        [4, 3, 5],
#        [7, 6, 8]])

# 二维数组交换两列
arr = np.arange(9).reshape(3,3)	# 原来的数组
arr[[1, 0, 2], :]
# array([[3, 4, 5],
#        [0, 1, 2],
#        [6, 7, 8]])

# 二维数组每行逆序输出
arr = np.arange(9).reshape(3,3)  # 原来的数组
arr[::-1]
# array([[6, 7, 8],
#        [3, 4, 5],
#        [0, 1, 2]])

# 二维数组每列逆序输出
arr = np.arange(9).reshape(3,3)  # 原来的数组
arr[::-1]
# array([[2, 1, 0],
#        [5, 4, 3],
#        [8, 7, 6]])

# 创建一个二维数组，数值区间为5-10
# 方法一
rand_arr = np.random.randint(low=5, high=10, size=(5,3)) + np.random.random((5,3))
# 方法二
rand_arr = np.random.uniform(5,10, size=(5,3))
rand_arr
# array([[ 6.75709583,  7.4662972 ,  7.2395679 ],
#        [ 9.7802112 ,  6.70830431,  6.76776651],
#        [ 6.35564151,  7.8632512 ,  8.53040631],
#        [ 7.99457763,  7.85526309,  8.96584465],
#        [ 6.18621991,  8.14214732,  7.31162037]])

21-30

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
import numpy as np

# 输出数组时，控制输出精度
rand_arr = np.random.random([3,3]) / 1e3  # 原始数组
rand_arr
# array([[  5.43404942e-04,   2.78369385e-04,   4.24517591e-04],
#        [  8.44776132e-04,   4.71885619e-06,   1.21569121e-04],
#        [  6.70749085e-04,   8.25852755e-04,   1.36706590e-04]])
# 第一个操作，指定小数点后3位数
np.set_printoptions(precision=3)
rand_arr
# array([[  5.434e-04,   2.784e-04,   4.245e-04],
#        [  8.448e-04,   4.719e-06,   1.216e-04],
#        [  6.707e-04,   8.259e-04,   1.367e-04]])
# 第二个操作，避免输出e-04，并指定小数点后6位数
np.set_printoptions(suppress=True, precision=6)
# array([[ 0.000543,  0.000278,  0.000425],
#        [ 0.000845,  0.000005,  0.000122],
#        [ 0.000671,  0.000826,  0.000137]])

# 含有很多元素的数组
a = np.arange(15)	# 原始数组
a
# array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14])
# 设定显示元素个数的阈值
np.set_printoptions(threshold=6)
a
# array([ 0,  1,  2, ..., 12, 13, 14])
# 设定不隐藏元素输出
np.set_printoptions(threshold=np.nan)
a
# array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14])

# 导入网站数据
url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data'
iris = np.genfromtxt(url, delimiter=',', dtype='object')
names = ('sepallength', 'sepalwidth', 'petallength', 'petalwidth', 'species')
iris[:3]
# array([[b'5.1', b'3.5', b'1.4', b'0.2', b'Iris-setosa'],
#        [b'4.9', b'3.0', b'1.4', b'0.2', b'Iris-setosa'],
#        [b'4.7', b'3.2', b'1.3', b'0.2', b'Iris-setosa']], dtype=object)

# 提取某一列
species = np.array([row[4] for row in iris])
species
# array([b'Iris-setosa', b'Iris-setosa', b'Iris-setosa', b'Iris-setosa',
#        b'Iris-setosa'],
#       dtype='|S15')

# 提取多列
# 方法一，在导入的时候指定列
iris = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0, 1, 2, 3])
# 方法二，使用上面的方法
iris = np.array([row.tolist()[:4] for row in iris], dtype='float')
# array([[ 5.1,  3.5,  1.4,  0.2],
#        [ 4.9,  3. ,  1.4,  0.2],
#        [ 4.7,  3.2,  1.3,  0.2],
#        [ 4.6,  3.1,  1.5,  0.2]])

# 计算平均值、方差、中位数
sepallength = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0])
mu, med, sd = np.mean(sepallength), np.median(sepallength), np.std(sepallength)
# 5.84333333333 5.8 0.825301291785

# 0-1化数组
sepallength = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0])
Smax, Smin = sepallength.max(), sepallength.min()
# 方法一：
S = (sepallength - Smin)/(Smax - Smin)
# 方法二：
S = (sepallength - Smin)/sepallength.ptp()
# array([ 0.222222,  0.166667,  0.111111, ...,  0.611111,  0.527778,
#         0.444444])

# softmax数组
sepallength = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0])
def softmax(x):
    e_x = np.exp(x - np.max(x))
    return e_x / e_x.sum(axis=0)

softmax(sepallength)
# array([ 0.00222 ,  0.001817,  0.001488, ...,  0.009001,  0.006668,  0.00494 ])

31-40

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
import numpy as np

url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data'

# 提取数据中百分位数
sepallength = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0])
# 提取数据中5%, 95%的数
np.percentile(sepallength, q=[5, 95])
# array([ 4.6  ,  7.255])

# 将数据中的缺失值替换为随机数
url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data'
iris_2d = np.genfromtxt(url, delimiter=',', dtype='object')
# 方法一
i, j = np.where(iris_2d)
np.random.seed(100)
iris_2d[np.random.choice((i), 20), np.random.choice((j), 20)] = np.nan
# 方法二
np.random.seed(100)
iris_2d[np.random.randint(150, size=20), np.random.randint(4, size=20)] = np.nan
iris_2d[:10]
# array([[b'5.1', b'3.5', b'1.4', b'0.2', b'Iris-setosa'],
#        [b'4.9', b'3.0', b'1.4', b'0.2', b'Iris-setosa'],
#        [b'4.7', b'3.2', b'1.3', b'0.2', b'Iris-setosa'],
#        [b'4.6', b'3.1', b'1.5', b'0.2', b'Iris-setosa'],
#        [b'5.0', b'3.6', b'1.4', b'0.2', b'Iris-setosa'],
#        [b'5.4', b'3.9', b'1.7', b'0.4', b'Iris-setosa'],
#        [b'4.6', b'3.4', b'1.4', b'0.3', b'Iris-setosa'],
#        [b'5.0', b'3.4', b'1.5', b'0.2', b'Iris-setosa'],
#        [b'4.4', nan, b'1.4', b'0.2', b'Iris-setosa'],
#        [b'4.9', b'3.1', b'1.5', b'0.1', b'Iris-setosa']], dtype=object)

# 统计数据中缺失值位置
url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data'
iris_2d = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3])
iris_2d[np.random.randint(150, size=20), np.random.randint(4, size=20)] = np.nan
# 统计缺失值个数
np.isnan(iris_2d[:, 0]).sum()
# 5
# 统计缺失值位置
np.where(np.isnan(iris_2d[:, 0]))
# (array([ 38,  80, 106, 113, 121], dtype=int64),)

# 二维数组多重条件筛选
condition = (iris_2d[:, 2] > 1.5) & (iris_2d[:, 0] < 5.0)
iris_2d[condition]
# array([[ 4.8,  3.4,  1.6,  0.2],
#        [ 4.8,  3.4,  1.9,  0.2],
#        [ 4.7,  3.2,  1.6,  0.2],
#        [ 4.8,  3.1,  1.6,  nan],
#        [ 4.9,  2.4,  3.3,  1. ]])

# 删除有缺失值的行，注意~
any_nan_in_row = np.array([~np.any(np.isnan(row)) for row in iris_2d])
# 未删除前
iris_2d.shape
# (150, 4)
# 删除后
iris_2d[any_nan_in_row].shape()
# (130, 4)

# 计算数组间关联度
iris = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3])
np.corrcoef(iris[:, 0], iris[:, 2])[0, 1]
# 0.87175415730487182

# 检查数组中是否含有缺失值
np.isnan(iris_2d).any()
# True

# 替换数组中的缺失值为0
iris_2d[np.isnan(iris_2d)] = 0
# 替换前
np.isnan(iris_2d).any()
# True
# 替换后
np.isnan(iris_2d).any()
# False

# 统计数组中非重复数值的情况
np.unique(iris, return_counts=True)
# 上面的数组的表示数字，下面的数组表示出现次数
# (array([ 0.1,  0.2,  0.3,  0.4,  0.5,  0.6,  1. ,  1.1,  1.2,  1.3,  1.4,
#          1.5,  1.6,  1.7,  1.8,  1.9,  2. ,  2.1,  2.2,  2.3,  2.4,  2.5,
#          2.6,  2.7,  2.8,  2.9,  3. ,  3.1,  3.2,  3.3,  3.4,  3.5,  3.6,
#          3.7,  3.8,  3.9,  4. ,  4.1,  4.2,  4.3,  4.4,  4.5,  4.6,  4.7,
#          4.8,  4.9,  5. ,  5.1,  5.2,  5.3,  5.4,  5.5,  5.6,  5.7,  5.8,
#          5.9,  6. ,  6.1,  6.2,  6.3,  6.4,  6.5,  6.6,  6.7,  6.8,  6.9,
#          7. ,  7.1,  7.2,  7.3,  7.4,  7.6,  7.7,  7.9]),
#  array([ 6, 28,  7,  7,  1,  1,  8,  4,  7, 20, 20, 26, 11,  6, 12,  7,  7,
#          6,  6, 12,  6, 11,  5,  9, 14, 10, 27, 12, 13,  8, 12,  8,  4,  4,
#          7,  5,  6,  4,  5,  3,  8,  9,  7,  7,  9, 11, 14, 17,  6,  3,  8,
#         10, 12, 11, 10,  5,  8,  9,  4, 10,  8,  5,  3, 10,  3,  5,  1,  1,
#          3,  1,  1,  1,  4,  1], dtype=int64))

# 将数组中处于相应区间的数换位文本
# 将数组中数组位于[0, 3, 5, 10]区间内的转换为对应的索引
petal_length_bin = np.digitize(iris[:, 2].astype('float'), [0, 3, 5, 10])
label_map = {1: 'small', 2: 'medium', 3: 'large', 4: np.nan}
petal_length_cat = [label_map[x] for x in petal_length_bin]
petal_length_cat[:4]
# ['small', 'small', 'small', 'small']

41-50

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
import numpy as np

url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data'

# 在已经存在的数组中，多加一列
iris_2d = np.genfromtxt(url, delimiter=',', dtype='object')
# 计算
sepallength = iris_2d[:, 0].astype('float')
petallength = iris_2d[:, 2].astype('float')
volume = (np.pi * petallength * (sepallength**2))/3
volume = volume[:, np.newaxis]
# 横向堆叠
out = np.hstack([iris_2d, volume])
out[:4]
# array([[b'5.1', b'3.5', b'1.4', b'0.2', b'Iris-setosa', 38.13265162927291],
#        [b'4.9', b'3.0', b'1.4', b'0.2', b'Iris-setosa', 35.200498485922445],
#        [b'4.7', b'3.2', b'1.3', b'0.2', b'Iris-setosa', 30.0723720777127],
#        [b'4.6', b'3.1', b'1.5', b'0.2', b'Iris-setosa', 33.238050274980004]], dtype=object)

# 对数组进行概率抽样
iris = np.genfromtxt(url, delimiter=',', dtype='object')
species = iris[:, 4]
# 方法一：生成概率抽样
np.random.seed(100)
a = np.array(['Iris-setosa', 'Iris-versicolor', 'Iris-virginica'])
species_out = np.random.choice(a, 150, p=[0.5, 0.25, 0.25])
# 方法二：直接概率抽样
np.random.seed(100)
probs = np.r_[np.linspace(0, 0.500, num=50), np.linspace(0.501, .750, num=50), np.linspace(.751, 1.0, num=50)]
index = np.searchsorted(probs, np.random.random(150))
species_out = species[index]
np.unique(species_out, return_counts=True)
# (array([b'Iris-setosa', b'Iris-versicolor', b'Iris-virginica'], dtype=object), array([77, 37, 36], dtype=int64))

# 对筛选后的数组，取出第二大的数
petal_len_setosa = iris[iris[:, 4] == b'Iris-setosa', [2]].astype('float')
np.unique(np.sort(petal_len_setosa))[-2]
# 1.7

# 对二维数组按照其中某一列进行排序
# 按照第一列升序
iris[iris[:,0].argsort()][:20]
# array([[b'4.3', b'3.0', b'1.1', b'0.1', b'Iris-setosa'],
#        [b'4.4', b'3.2', b'1.3', b'0.2', b'Iris-setosa'],
#        [b'4.4', b'3.0', b'1.3', b'0.2', b'Iris-setosa'],
#        [b'4.4', b'2.9', b'1.4', b'0.2', b'Iris-setosa'],
#        [b'4.5', b'2.3', b'1.3', b'0.3', b'Iris-setosa'],
#        [b'4.6', b'3.6', b'1.0', b'0.2', b'Iris-setosa'],
#        [b'4.6', b'3.1', b'1.5', b'0.2', b'Iris-setosa'],
#        [b'4.6', b'3.4', b'1.4', b'0.3', b'Iris-setosa'],
#        [b'4.6', b'3.2', b'1.4', b'0.2', b'Iris-setosa'],
#        [b'4.7', b'3.2', b'1.3', b'0.2', b'Iris-setosa'],
#        [b'4.7', b'3.2', b'1.6', b'0.2', b'Iris-setosa'],
#        [b'4.8', b'3.0', b'1.4', b'0.1', b'Iris-setosa'],
#        [b'4.8', b'3.0', b'1.4', b'0.3', b'Iris-setosa'],
#        [b'4.8', b'3.4', b'1.9', b'0.2', b'Iris-setosa'],
#        [b'4.8', b'3.4', b'1.6', b'0.2', b'Iris-setosa'],
#        [b'4.8', b'3.1', b'1.6', b'0.2', b'Iris-setosa'],
#        [b'4.9', b'2.4', b'3.3', b'1.0', b'Iris-versicolor'],
#        [b'4.9', b'2.5', b'4.5', b'1.7', b'Iris-virginica'],
#        [b'4.9', b'3.1', b'1.5', b'0.1', b'Iris-setosa'],
#        [b'4.9', b'3.1', b'1.5', b'0.1', b'Iris-setosa']], dtype=object)

# 统计出现次数最多的数
vals, counts = np.unique(iris[:, 3], return_counts=True)
vals[np.argmax(counts)]
# b'0.2'

# 找出数组中第一次出现比给定数字大的值的缩影
# 这里设定要找比1.0大的值
np.argmax(iris[:, 3].astype('float') > 1.0)
# 50

# 给定一个区间，如果数字不在这个区间内，就替换为区间的最小/最大值
a = np.random.uniform(1,50, 20)  # 原始数组
# array([ 22.153505  ,  47.06146116,  41.06481956,  17.46948556,
#          9.59511223,  19.26877027,   1.27873686,  13.36889132,
#         39.98746292,   1.74749359,  30.34332547,  30.58642241,
#          6.15223659,  19.7152288 ,   2.78732677,  44.63016661,
#         49.06512199,   3.93715745,  44.63675129,  29.26817347])
# 方法一：
np.clip(a, a_min=10, a_max=30)
# 方法二：
np.where(a < 10, 10, np.where(a > 30, 30, a))
# array([ 22.153505  ,  30.        ,  30.        ,  17.46948556,
#         10.        ,  19.26877027,  10.        ,  13.36889132,
#         30.        ,  10.        ,  30.        ,  30.        ,
#         10.        ,  19.7152288 ,  10.        ,  30.        ,
#         30.        ,  10.        ,  30.        ,  29.26817347])

# 统计前n大的元素位置
a = np.random.uniform(1,50, 20)  # 原始数组
a
# array([ 22.153505  ,  47.06146116,  41.06481956,  17.46948556,
#          9.59511223,  19.26877027,   1.27873686,  13.36889132,
#         39.98746292,   1.74749359,  30.34332547,  30.58642241,
#          6.15223659,  19.7152288 ,   2.78732677,  44.63016661,
#         49.06512199,   3.93715745,  44.63675129,  29.26817347])
# 方法一：
a.argsort()[-5:]
# array([ 2, 15, 18,  1, 16], dtype=int64)
# 方法二：
np.argpartition(-a, 5)[:5]
# array([16,  1, 18, 15,  2], dtype=int64)
# 获取前n大的元素
a[a.argsort()][-5:]
# 方法1:
a[a.argsort()][-5:]
# 方法2:
np.sort(a)[-5:]
# 方法3:
np.partition(a, kth=-5)[-5:]
# 前三个输出都是
# array([ 41.06481956,  44.63016661,  47.06146116,  44.63675129,  49.06512199])
# 方法4:
a[np.argpartition(-a, 5)][:5]
# array([ 49.06512199,  47.06146116,  44.63675129,  44.63016661,  41.06481956])


# 计算有唯一值的行数,没看懂
arr = np.random.randint(1,11,size=(6, 10))
arr
# array([[ 2,  6,  5,  3,  9,  4,  6,  1, 10,  4],
#        [ 7,  4,  5,  8,  7,  4, 10,  1,  5,  5],
#        [ 6,  8,  7,  7,  3,  5,  3,  8,  2,  7],
#        [ 7,  1,  8,  3,  4,  6,  5,  3,  5,  4],
#        [ 8, 10,  1,  1,  6, 10,  7,  7,  6,  7],
#        [ 5,  8,  4, 10,  3,  4,  9,  8,  2,  6]])
def counts_of_all_values_rowwise(arr2d):
    num_counts_array = [np.unique(row, return_counts=True) for row in arr2d]
    return [[int(b[a==i]) if i in a else 0 for i in np.unique(arr2d)] for a, b in num_counts_array]
counts_of_all_values_rowwise(arr)
# [[1, 1, 1, 2, 1, 2, 0, 0, 1, 1],
#  [1, 0, 0, 2, 3, 0, 2, 1, 0, 1],
#  [0, 1, 2, 0, 1, 1, 3, 2, 0, 0],
#  [1, 0, 2, 2, 2, 1, 1, 1, 0, 0],
#  [2, 0, 0, 0, 0, 2, 3, 1, 0, 2],
#  [0, 1, 1, 2, 1, 1, 0, 2, 1, 1]]

# 将多维数组转换为一维数组
arr1 = np.arange(3)
arr2 = np.arange(3,7)
arr3 = np.arange(7,10)
array_of_arrays = np.array([arr1, arr2, arr3])
array_of_arrays      # 原始数组
# array([array([0, 1, 2]), array([3, 4, 5, 6]), array([7, 8, 9])], dtype=object)
arr_2d = np.concatenate(array_of_arrays)
arr_2d
# array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

51-60

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
import numpy as np

# 唯一编码位置
np.random.seed(101)
arr = np.random.randint(1,4, size=6)
arr
# array([2, 3, 2, 2, 2, 1])
# 返回数组是3x6,因为数组长度为6，最大数为1。1对应的是[1,0,0,],2对应的是[0,1,0],3对应的是[0,0,1]
# 方法一
def one_hot_encodings(arr):
    uniqs = np.unique(arr)
    out = np.zeros((arr.shape[0], uniqs.shape[0]))
    for i, k in enumerate(arr):
        out[i, k-1] = 1
    return out
one_hot_encodings(arr)
# array([[ 0.,  1.,  0.],
#        [ 0.,  0.,  1.],
#        [ 0.,  1.,  0.],
#        [ 0.,  1.,  0.],
#        [ 0.,  1.,  0.],
#        [ 1.,  0.,  0.]])
# 方法二
(arr[:, None] == np.unique(arr)).view(np.int8)
# array([[0, 1, 0],
#        [0, 0, 1],
#        [0, 1, 0],
#        [0, 1, 0],
#        [0, 1, 0],
#        [1, 0, 0]], dtype=int8)

# 获取由变量名分组的行号
url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data'
species = np.genfromtxt(url, delimiter=',', dtype='str', usecols=4)
np.random.seed(100)
species_small = np.sort(np.random.choice(species, size=20))
species_small
# array(['Iris-setosa', 'Iris-setosa', 'Iris-setosa', 'Iris-setosa',
#        'Iris-setosa', 'Iris-versicolor', 'Iris-versicolor',
#        'Iris-versicolor', 'Iris-versicolor', 'Iris-versicolor',
#        'Iris-versicolor', 'Iris-versicolor', 'Iris-versicolor',
#        'Iris-versicolor', 'Iris-virginica', 'Iris-virginica',
#        'Iris-virginica', 'Iris-virginica', 'Iris-virginica',
#        'Iris-virginica'],
#       dtype='<U15')

[i for val in np.unique(species_small) for i, grp in enumerate(species_small[species_small==val])]
[0, 1, 2, 3, 4, 0, 1, 2, 3, 4, 5, 6, 7, 8, 0, 1, 2, 3, 4, 5]

# 根据已知分类组名创建索引
np.random.seed(100)
species_small = np.sort(np.random.choice(species, size=20))
species_small
# array(['Iris-setosa', 'Iris-setosa', 'Iris-setosa', 'Iris-setosa',
#        'Iris-setosa', 'Iris-setosa', 'Iris-versicolor', 'Iris-versicolor',
#        'Iris-versicolor', 'Iris-versicolor', 'Iris-versicolor',
#        'Iris-versicolor', 'Iris-virginica', 'Iris-virginica',
#        'Iris-virginica', 'Iris-virginica', 'Iris-virginica',
#        'Iris-virginica', 'Iris-virginica', 'Iris-virginica'],
#       dtype='<U15')
# 方法一，用列表解析式
output = [np.argwhere(np.unique(species_small) == s).tolist()[0][0] for val in np.unique(species_small) for s in species_small[species_small==val]]
# 方法二，不用列表解析式直接循环
output = []
uniqs = np.unique(species_small)
for val in uniqs:  # uniq values in group
    for s in species_small[species_small==val]:  # each element in group
        groupid = np.argwhere(uniqs == s).tolist()[0][0]  # groupid
        output.append(groupid)
output
# [0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2]

# 数值排序，输出对应的索引值
np.random.seed(10)
a = np.random.randint(20, size=10)
a
# [ 9  4 15  0 17 16 17  8  9  0]
a.argsort().argsort()
# [4 2 6 0 8 7 9 3 5 1]

# 多维数组排序，输出对应的索引值
np.random.seed(10)
a = np.random.randint(20, size=[2,5])
a
# [[ 9  4 15  0 17]
#  [16 17  8  9  0]]
a.ravel().argsort().argsort().reshape(a.shape)
# [[ 9  4 15  0 17]
#  [16 17  8  9  0]]
# [[4 2 6 0 8]
#  [7 9 3 5 1]]

# 输出多维数组每一行的最大值
np.random.seed(100)
a = np.random.randint(1,10, [5,3])
a
# array([[9, 9, 4],
#        [8, 8, 1],
#        [5, 3, 6],
#        [3, 3, 3],
#        [2, 1, 9]])
# 方法一
np.amax(a, axis=1)
# 方法二
np.apply_along_axis(np.max, arr=a, axis=1)
# array([9, 8, 6, 3, 9])

# 输出多维数组每行最小值比最大值
np.random.seed(100)
a = np.random.randint(1,10, [5,3])
a
# array([[9, 9, 4],
#        [8, 8, 1],
#        [5, 3, 6],
#        [3, 3, 3],
#        [2, 1, 9]])
np.apply_along_axis(lambda x: np.min(x)/np.max(x), arr=a, axis=1)
# array([ 0.44444444,  0.125     ,  0.5       ,  1.        ,  0.11111111])

# 找出数组中重复的数值，输出布尔型
np.random.seed(100)
a = np.random.randint(0, 5, 10)
a
# [0 0 3 0 2 4 2 2 2 2]
out = np.full(a.shape[0], True)
unique_positions = np.unique(a, return_index=True)[1]
out[unique_positions] = False
out
# [False  True False  True False False  True  True  True  True]

# 输出多维数组每一分类的平均值
iris = np.genfromtxt(url, delimiter=',', dtype='object')
names = ('sepallength', 'sepalwidth', 'petallength', 'petalwidth', 'species')
numeric_column = iris[:, 1].astype('float')  # sepalwidth
grouping_column = iris[:, 4]  # species

# 方法一：列表解析
output = [[group_val, numeric_column[grouping_column==group_val].mean()] for group_val in np.unique(grouping_column)]
# 方法二：不用列表解析直接循环
output = []
for group_val in np.unique(grouping_column):
    output.append([group_val, numeric_column[grouping_column==group_val].mean()])
output
# [[b'Iris-setosa', 3.418],
#  [b'Iris-versicolor', 2.770],
#  [b'Iris-virginica', 2.974]]

# 将图像转换为数组存储，并再转换出图像显示
from io import BytesIO
from PIL import Image
import PIL, requests

URL = 'https://upload.wikimedia.org/wikipedia/commons/8/8b/Denali_Mt_McKinley.jpg'
response = requests.get(URL)
I = Image.open(BytesIO(response.content))
I = I.resize([150,150])
arr = np.asarray(I)
im = PIL.Image.fromarray(np.uint8(arr))
Image.Image.show(im)

61-70

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
import numpy as np

# 去除数组中所有的缺失值
a = np.array([1,2,3,np.nan,5,6,7,np.nan])
a[~np.isnan(a)]
array([ 1.,  2.,  3.,  5.,  6.,  7.])

# 计算两个数组间的欧几里得
a = np.array([1,2,3,4,5])
b = np.array([4,5,6,7,8])
dist = np.linalg.norm(a-b)
dist
# 6.7082039324993694

# 找出数组中最大的两个数
a = np.array([1, 3, 7, 1, 2, 6, 0, 1])
doublediff = np.diff(np.sign(np.diff(a)))
peak_locations = np.where(doublediff == -2)[0] + 1
peak_locations
# array([2, 5])

# 二维数组数组数值每项减去一维数组对应的行号
a_2d = np.array([[3,3,3],[4,4,4],[5,5,5]])
b_1d = np.array([1,2,3])
a_2d - b_1d[:,None]
# [[2 2 2]
#  [2 2 2]
#  [2 2 2]]

# 数组中重复的第n次的数的索引
x = np.array([1, 2, 1, 1, 3, 4, 3, 1, 1, 2, 1, 1, 2])
n = 5
# 方法一：列表解析
[i for i, v in enumerate(x) if v == 1][n-1]
# 方法二：numpy方法
np.where(x == 1)[0][n-1]
# 8

# 将`datetime64`格式转换为`datetime`
dt64 = np.datetime64('2018-02-25 22:10:10')
# 方法一：利用datetime模块
from datetime import datetime
dt64.tolist()
# 方法二：numpy直接转换
dt64.astype(datetime)
datetime.datetime(2018, 2, 25, 22, 10, 10)

# 这个没看懂。。
# 如何计算numpy数组的移动平均值(累加和)
np.random.seed(100)
Z = np.random.randint(10, size=10)
Z
# [8 8 3 7 7 0 4 2 5 2]
# 方法一: np.cumsum
def moving_average(a, n=3) :
    ret = np.cumsum(a, dtype=float)
    ret[n:] = ret[n:] - ret[:-n]
    return ret[n - 1:] / n
moving_average(Z, n=3).round(2)
# 方法二: np.convolve
np.convolve(Z, np.ones(3)/3, mode='valid')
# [ 6.33  6.    5.67  4.67  3.67  2.    3.67  3.  ]

# 创建一个数组，只给出起始值、步长、数组长度
length = 10
start = 5
step = 3
def seq(start, length, step):
    end = start + (step*length)
    return np.arange(start, end, step)
seq(start, length, step)
# array([ 5,  8, 11, 14, 17, 20, 23, 26, 29, 32])

# numpy数组填补不完整的日期
dates = np.arange(np.datetime64('2018-02-01'), np.datetime64('2018-02-25'), 2)
dates
# ['2018-02-01' '2018-02-03' '2018-02-05' '2018-02-07' '2018-02-09'
#  '2018-02-11' '2018-02-13' '2018-02-15' '2018-02-17' '2018-02-19'
#  '2018-02-21' '2018-02-23']
# 方法一: 列表解析
filled_in = np.array([np.arange(date, (date+d)) for date, d in zip(dates, np.diff(dates))]).reshape(-1)
output = np.hstack([filled_in, dates[-1]])
# 方法二: 不用列表解析循环
out = []
for date, d in zip(dates, np.diff(dates)):
    out.append(np.arange(date, (date+d)))
filled_in = np.array(out).reshape(-1)
output = np.hstack([filled_in, dates[-1]])
output
# array(['2018-02-01', '2018-02-02', '2018-02-03', '2018-02-04',
#        '2018-02-05', '2018-02-06', '2018-02-07', '2018-02-08',
#        '2018-02-09', '2018-02-10', '2018-02-11', '2018-02-12',
#        '2018-02-13', '2018-02-14', '2018-02-15', '2018-02-16',
#        '2018-02-17', '2018-02-18', '2018-02-19', '2018-02-20',
#        '2018-02-21', '2018-02-22', '2018-02-23'], dtype='datetime64[D]')

# 给定一维数组，变形为二维数组（给定每维长度）
arr = np.arange(15)
arr
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14])
def gen_strides(arr, stride_len=5, window_len=5):
    n_strides = (a.size-window_len)//stride_len + 1
    return np.array([a[s:(s+window_len)] for s in np.arange(0, n_strides*stride_len, stride_len)])
gen_strides(np.arange(15), stride_len=2, window_len=4)
# [[ 0  1  2  3]
#  [ 2  3  4  5]
#  [ 4  5  6  7]
#  [ 6  7  8  9]
#  [ 8  9 10 11]
#  [10 11 12 13]]