numpy

NumPy及向量化

In [100]:

测试在jupyter中进行：

import numpy as np

1. 1 创建Array

my_list = [1, 2, 3]

x = np.array(my_list)

print('Array: ', x)

Array:  [1 2 3]

x.shape

(3,)

x.ndim

1

np.array([1, 2, 3]) - np.array([4, 5, 6])

 array([-3, -3, -3])

m = np.array([[1, 2, 3], [4, 5, 6]])

print(m)

print('shape: ', m.shape)

[[1 2 3]
 [4 5 6]]
shape:  (2, 3)

 

n = np.arange(0, 30, 2)

print(n)

[ 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28]

n = n.reshape(3, 5)
print('reshape后: ')

print(n)

reshape后: 
[[ 0  2  4  6  8]
 [10 12 14 16 18]
 [20 22 24 26 28]]

 

print('ones:\n', np.ones((3, 2)))

print('zeros:\n', np.zeros((3, 2)))

print('eye:\n', np.eye(3))

print('diag:\n', np.diag(my_list))

ones:
 [[ 1.  1.]
 [ 1.  1.]
 [ 1.  1.]]
zeros:
 [[ 0.  0.]
 [ 0.  0.]
 [ 0.  0.]]
eye:
 [[ 1.  0.  0.]
 [ 0.  1.  0.]
 [ 0.  0.  1.]]
diag:
 [[1 0 0]
 [0 2 0]
 [0 0 3]]

 

print('*操作：\n', np.array([1, 2, 3] * 3))

print('repeat：\n', np.repeat([1, 2, 3], 3))

*操作：
 [1 2 3 1 2 3 1 2 3]
repeat：
 [1 1 1 2 2 2 3 3 3]

 

p1 = np.ones((3, 3))

p2 = np.arange(9).reshape(3, 3)

print('纵向叠加: \n', np.vstack((p1, p2))

print('横向叠加: \n', np.hstack((p1, p2)))

纵向叠加: 
 [[ 1.  1.  1.]
 [ 1.  1.  1.]
 [ 1.  1.  1.]
 [ 0.  1.  2.]
 [ 3.  4.  5.]
 [ 6.  7.  8.]]
横向叠加: 
 [[ 1.  1.  1.  0.  1.  2.]
 [ 1.  1.  1.  3.  4.  5.]
 [ 1.  1.  1.  6.  7.  8.]]

1.2. Array操作

print('p1: \n', p1)

print('p2: \n', p2)

print('p1 + p2 = \n', p1 + p2）

print('p1 * p2 = \n', p1 * p2)

print('p2^2 = \n', p2 ** 2)
print('p1.p2 = \n', p1.dot(p2))

p1: 
 [[ 1.  1.  1.]
 [ 1.  1.  1.]
 [ 1.  1.  1.]]
p2: 
 [[0 1 2]
 [3 4 5]
 [6 7 8]]
p1 + p2 = 
 [[ 1.  2.  3.]
 [ 4.  5.  6.]
 [ 7.  8.  9.]]
p1 * p2 = 
 [[ 0.  1.  2.]
 [ 3.  4.  5.]
 [ 6.  7.  8.]]
p2^2 = 
 [[ 0  1  4]
 [ 9 16 25]
 [36 49 64]]
p1.p2 = 
 [[  9.  12.  15.]
 [  9.  12.  15.]
 [  9.  12.  15.]]

p3 = np.arange(6).reshape(2, 3)

print('p3形状: ', p3.shape)

print(p3)

p4 = p3.T

print('转置后p3形状: ', p4.shape)

print(p4)

p3形状:  (2, 3)
[[0 1 2]
 [3 4 5]]
转置后p3形状:  (3, 2)
[[0 3]
 [1 4]
 [2 5]]

p3.reshape(-1, 1)

 

array([[0],
       [1],
       [2],
       [3],
       [4],
       [5]])

print('p3数据类型:', p3.dtype)

print(p3)

p5 = p3.astype('float')

print('p5数据类型:', p5.dtype)

print(p5)

 

p3数据类型: int32
[[0 1 2]
 [3 4 5]]
p5数据类型: float64
[[ 0.  1.  2.]
 [ 3.  4.  5.]]

a = np.array([-4, -2, 1, 3, 5])
# print('sum: ', a.sum())
# print('min: ', a.min())
# print('max: ', a.max())
# print('mean: ', a.mean())
# print('std: ', a.std())
print('argmax: ', a.argmax())
print('argmin: ', a.argmin())

argmax:  4
argmin:  0

1.3. 索引与切片

# 一维array

s = np.arange(13) ** 2

print('s: ', s)

print('s[0]: ', s[0])

print('s[4]: ', s[4])

print('s[0:3]: ', s[0:3])

print('s[[0, 2, 4]]: ', s[[0, 2, 4]])

s:  [  0   1   4   9  16  25  36  49  64  81 100 121 144]
s[0]:  0
s[4]:  16
s[0:3]:  [0 1 4]
s[[0, 2, 4]]:  [ 0  4 16]

 

# 二维array
r = np.arange(36).reshape((6, 6))
print('r: \n', r)
print('r[2, 2]: \n', r[2, 2])
print('r[3, 3:6]: \n', r[3:5, 3:6])

r: 
 [[ 0  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]]
r[2, 2]: 
 14
r[3, 3:6]: 
 [[21 22 23]
 [27 28 29]]

r > 30

array([[False, False, False, False, False, False],
       [False, False, False, False, False, False],
       [False, False, False, False, False, False],
       [False, False, False, False, False, False],
       [False, False, False, False, False, False],
       [False,  True,  True,  True,  True,  True]], dtype=bool)

# # 过滤
# print(r[r > 30])

# 将大于30的数赋值为30
r[r > 30] = 30
print(r)

[[ 0  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 30 30 30 30 30]]

# copy()操作
r2 = r[:3, :3]
print(r2)

[[ 0  1  2]
 [ 6  7  8]
 [12 13 14]]

# 将r2内容设置为0
r2[:] = 0

# 查看r的内容
print(r

[[ 0  0  0  3  4  5]
 [ 0  0  0  9 10 11]
 [ 0  0  0 15 16 17]
 [18 19 20 21 22 23]
 [24 25 26 27 28 29]
 [30 30 30 30 30 30]]

r3 = r.copy()
r3[:] = 0
print(r).

[[ 0  0  0  3  4  5]
 [ 0  0  0  9 10 11]
 [ 0  0  0 15 16 17]
 [18 19 20 21 22 23]
 [24 25 26 27 28 29]
 [30 30 30 30 30 30]]

1.4. 遍历 Array

t = np.random.randint(0, 10, (4, 3))
print(t)

[[3 4 4]
 [9 0 9]
 [8 8 8]
 [6 4 3]]

for row in t:
    print(row)

[3 4 4]
[9 0 9]
[8 8 8]
[6 4 3]

# 使用enumerate()
for i, row in enumerate(t):
    print('row {} is {}'.format(i, row))

row 0 is [3 4 4]
row 1 is [9 0 9]
row 2 is [8 8 8]
row 3 is [6 4 3]

t2 = t ** 2
print(t2)

row 0 is [3 4 4]
row 1 is [9 0 9]
row 2 is [8 8 8]
row 3 is [6 4 3]

# 使用zip对两个array进行遍历计算
for i, j in zip(t, t2):
    print('{} + {} = {}'.format(i, j, i + j))

[3 4 4] + [ 9 16 16] = [12 20 20]
[9 0 9] + [81  0 81] = [90  0 90]
[8 8 8] + [64 64 64] = [72 72 72]
[6 4 3] + [36 16  9] = [42 20 12]