In [1]:
import numpy as np
In [2]:
# 算术运算
a = np.random.random(10000)*10
b = np.random.random(10000)*10
a
b
Out[2]:
array([9.66314537, 9.56406383, 4.43887791, ..., 7.85109353, 6.34951408,
2.53352231])
In [3]:
a
Out[3]:
array([9.34912059, 1.72634859, 6.30775754, ..., 8.32242753, 8.251723 ,
7.08478662])
In [4]:
b
Out[4]:
array([9.66314537, 9.56406383, 4.43887791, ..., 7.85109353, 6.34951408,
2.53352231])
In [5]:
# 既可以使用运算符也可以使用np算术函数
a + b
Out[5]:
array([19.01226595, 11.29041242, 10.74663546, ..., 16.17352106,
14.60123708, 9.61830893])
In [6]:
np.add(a,b) # 加
Out[6]:
array([19.01226595, 11.29041242, 10.74663546, ..., 16.17352106,
14.60123708, 9.61830893])
In [7]:
np.subtract(a,b) # 减
Out[7]:
array([-0.31402478, -7.83771524, 1.86887963, ..., 0.47133401,
1.90220891, 4.55126431])
In [8]:
np.multiply(a,b) # 乘
Out[8]:
array([90.34191128, 16.51090807, 27.99936564, ..., 65.34015693,
52.39443141, 17.94946495])
In [9]:
np.divide(a,b) # 除
Out[9]:
array([0.96750284, 0.18050367, 1.42102524, ..., 1.06003419, 1.29958338,
2.7964177 ])
In [10]:
# 数学函数-三角函数
ang = np.array([0,30,45,60,90,120,135,150,180])
rad = np.sin(ang*np.pi/180)
rad
Out[10]:
array([0.00000000e+00, 5.00000000e-01, 7.07106781e-01, 8.66025404e-01,
1.00000000e+00, 8.66025404e-01, 7.07106781e-01, 5.00000000e-01,
1.22464680e-16])
In [11]:
ang = np.array([0,30,45,60,90,120,135,150,180])
rad = np.cos(ang*np.pi/180)
rad
Out[11]:
array([ 1.00000000e+00, 8.66025404e-01, 7.07106781e-01, 5.00000000e-01,
6.12323400e-17, -5.00000000e-01, -7.07106781e-01, -8.66025404e-01,
-1.00000000e+00])
In [12]:
ang = np.array([0,30,45,60,90,120,135,150,180])
rad = np.tan(ang*np.pi/180)
rad
Out[12]:
array([ 0.00000000e+00, 5.77350269e-01, 1.00000000e+00, 1.73205081e+00,
1.63312394e+16, -1.73205081e+00, -1.00000000e+00, -5.77350269e-01,
-1.22464680e-16])
In [14]:
# 统计
d2 = np.array([1,2,3,4,5])
d2.mean()
Out[14]:
3.0
In [15]:
np.average(d2)
Out[15]:
3.0
In [16]:
np.average(d2,weights=np.array([2,2,1,1,1])) # 可以设置权重
Out[16]:
2.5714285714285716
In [17]:
np.average(d2,weights=np.array([1,1,1,2,3]))
Out[17]:
3.625
In [18]:
# 中位数
np.median(d2)
Out[18]:
3.0
In [19]:
d3 = np.array([1,2,3,4,5,6])
np.median(d3)
Out[19]:
3.5
In [20]:
# 方差
np.var(d3)
Out[20]:
2.9166666666666665
In [21]:
# 标准差
np.std(d3)
Out[21]:
1.707825127659933
In [22]:
# 聚合运算
a
Out[22]:
array([9.34912059, 1.72634859, 6.30775754, ..., 8.32242753, 8.251723 ,
7.08478662])
In [23]:
b
Out[23]:
array([9.66314537, 9.56406383, 4.43887791, ..., 7.85109353, 6.34951408,
2.53352231])
In [24]:
%time sum(a)
%time np.sum(a)
CPU times: total: 0 ns Wall time: 0 ns CPU times: total: 0 ns Wall time: 0 ns
Out[24]:
49855.5814326455
In [25]:
np.prod(d2)
Out[25]:
120
In [26]:
np.max(d2)
Out[26]:
5
In [27]:
# 筛选
d3
Out[27]:
array([1, 2, 3, 4, 5, 6])
In [28]:
np.where(d3>3) # 返回的索引
Out[28]:
(array([3, 4, 5], dtype=int64),)
In [29]:
np.where(d3<3) # 返回的索引
Out[29]:
(array([0, 1], dtype=int64),)
In [30]:
np.sort([5,4,2,7,2,1])
Out[30]:
array([1, 2, 2, 4, 5, 7])