Numpy数据运算

1.4-NumPy数组运算

 

 

1.4.1 算术运算¶

• add 加
• subtract 减
• multipie 乘
• divide 除

1.4.2 数学运算¶

• 三角函数: sin,cos,tan
• 算术运算

1.4.3 统计运算¶

• mean 均值
• average 均值
• var 方差
• std 标准差

1.4.4 聚合运算¶

• 性能区别: np聚合函数和py内置函数
• sum 累加
• prod 累乘
• max 最大值
• min 最小值
• 统计运算

1.4.5 筛选和排序¶

• where
• sort

 

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])

 

In [ ]: