NumPy学习

矩阵操作

基本操作

• 向量
• 类型
• 取值
• 判断
• 矩阵
• 维度
• 求值
• 比较
• 特定赋值
• 类型转换
• 求最值

import numpy as np

'向量'
v = np.array([1, 2, 3, 4])
print(v)
# [1 2 3 4]
'类型：会自动转化为同一类型'
print(v.dtype)
# int32
'求维度'
print(v.shape)
# (4,)
'向量取值'
print(v[1:3])
# [2 3]
'向量判断'
print(v == 2)
# [False  True False False]
t = (v == 2)
print(v[t])
# [2]

'矩阵'
m = np.array([
    [1, 2, 3],
    [4, 5, 6]
])
print(m)
# [[1 2 3]
#  [4 5 6]]
print(m.shape)
# (2, 3)
'取值：index从0开始'
print(m[1, 0])
# 4
'矩阵求值：冒号表示所有行即占位符，以逗号分割表示行列'
print(m[:, 1])
# [2 5]
print(m[:1, 0:1])
# [[1]]
'矩阵判断'
print(m == 3)
# [[False False  True]
#  [False False False]]
t = (m[:, 2] == 3)
print(m[t, :])
# [[1 2 3]]

'向量比较'
v = np.array([5, 10, 15, 20])
t = (v == 10) & (v == 5)
print(t)
# [False False False False]
v = np.array([5, 10, 15, 20])
t = (v == 10) | (v == 5)
print(t)
# [ True  True False False]
'特定赋值'
v[t] = 66
print(v)
# [66 66 15 20]

'矩阵特定赋值'
m = np.array([
    [5, 10, 15],
    [20, 25, 30],
    [35, 40, 45]
])
t = m[:, 1] == 25
print(t)
# [False  True False]
m[t, 1] = -1
print(m)
# [[ 5 10 15]
#  [20 -1 30]
#  [35 40 45]]

'向量的类型转换'
v = np.array(['1', '2', '3'])
print(v.dtype)
# <U1
print(v)
# ['1' '2' '3']
v = v.astype(float)
print(v.dtype)
# float64
print(v)
# [1. 2. 3.]

'向量求最值'
v = np.array([1, 2, 3, 4])
print(v.max(), v.min())
# 4 1

'矩阵求和：axis=0表示按列求和, axis=1表示按行求和'
m = np.array([
    [1, 2, 3],
    [3, 4, 5],
    [5, 6, 7]
])
print(m.sum(axis=1))
# [ 30  75 120]

矩阵操作

• 生成序列
• 序列转矩阵
• 维数
• 元素个数
• 零矩阵
• 一矩阵
• 随机矩阵
• 范围矩阵
• 矩阵加减
• 矩阵乘方
• 矩阵比较
• 矩阵内积
• 矩阵外积

import numpy as np

print(np.arange(15))
# [ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14]
a = np.arange(15).reshape(3, 5)
print(a)
# [[ 0  1  2  3  4]
#  [ 5  6  7  8  9]
#  [10 11 12 13 14]]
print(a.shape)
# (3, 5)
print(a.dtype.name)
# int32
'求维数'
print(a.ndim)
# 2
'求元素个数'
print(a.size)
# 15

print(np.zeros((3, 4)))
# [[0. 0. 0. 0.]
#  [0. 0. 0. 0.]
#  [0. 0. 0. 0.]]

print(np.ones((2, 3, 4), dtype=np.int32))
# [[[1 1 1 1]
#   [1 1 1 1]
#   [1 1 1 1]]
#
#  [[1 1 1 1]
#   [1 1 1 1]
#   [1 1 1 1]]]

print(np.arange(10, 30, 5))
# [10 15 20 25]

print(np.random.random((2, 3)))
# [[0.44358543 0.52329237 0.62330638]
#  [0.4742094  0.17395022 0.36088141]]

from numpy import pi

'从0开始到2*pi结束，一共10个数'
print(np.linspace(0, 2*pi, 10))
# [0.         0.6981317  1.3962634  2.0943951  2.7925268  3.4906585
#  4.1887902  4.88692191 5.58505361 6.28318531]

'矩阵运算'
a = np.array([20, 30, 40, 50])
b = np.arange(4)
print(a)
# [20 30 40 50]
print(b)
# [0 1 2 3]
c = a - b
print(c)
# [20 29 38 47]
c = c - 1
print(c)
# [19 28 37 46]
print(b ** 2)
# [0 1 4 9]
print(a < 35)
# [ True  True False False]

a = np.array([
    [1, 1],
    [0, 1]
])
b = np.array([
    [2, 0],
    [3, 4]
])
'矩阵内积：对应位置相乘'
print(a * b)
# [[2 0]
#  [0 4]]
'矩阵外积：矩阵乘法'
print(a.dot(b)) # 或print(np.dot(a, b))
# [[5 4]
#  [3 4]]

常用函数

• e次方
• 矩阵转序列
• 矩阵转置
• 拼接矩阵
• 切分矩阵
• 复制矩阵
• 求矩阵下标
• 矩阵整体变换
• 矩阵排序

import numpy as np

b = np.arange(3)
print(b)
# [0 1 2]
print(np.exp(b))
# [1.         2.71828183 7.3890561 ]
print(np.sqrt(b))
# [0.         1.         1.41421356]


a = np.floor(10 * np.random.random((3, 4)))
print(a)
# [[3. 4. 5. 9.]
#  [3. 5. 1. 8.]
#  [0. 4. 3. 7.]]
print(a.ravel())
# [3. 4. 5. 9. 3. 5. 1. 8. 0. 4. 3. 7.]
a.shape = (6, 2)
print(a)
# [[3. 4.]
#  [5. 9.]
#  [3. 5.]
#  [1. 8.]
#  [0. 4.]
#  [3. 7.]]
print(a.T)
# [[3. 5. 3. 1. 0. 3.]
#  [4. 9. 5. 8. 4. 7.]]


a = np.floor(10 * np.random.random((2, 2)))
b = np.floor(10 * np.random.random((2, 2)))
print(a)
# [[2. 0.]
#  [9. 6.]]
print(b)
# [[1. 5.]
#  [1. 8.]]
'拼接矩阵：vstack是垂直拼接，hstack是水平拼接'
print(np.vstack((a, b)))
# [[2. 0.]
#  [9. 6.]
#  [1. 5.]
#  [1. 8.]]


a = np.floor(10 * np.random.random((2, 12)))
print(a)
# [[3. 0. 4. 5. 5. 3. 7. 8. 2. 6. 9. 0.]
#  [2. 8. 2. 7. 7. 8. 6. 1. 1. 1. 9. 8.]]
'平均切分，同上分水平和垂直'
print(np.hsplit(a, 3))
# [array([[3., 0., 4., 5.],
#        [2., 8., 2., 7.]]), array([[5., 3., 7., 8.],
#        [7., 8., 6., 1.]]), array([[2., 6., 9., 0.],
#        [1., 1., 9., 8.]])]
'指定位置切分，同上分水平和垂直'
print(np.hsplit(a, (3, 4)))
# [array([[3., 0., 4.],
#        [2., 8., 2.]]), array([[5.],
#        [7.]]), array([[5., 3., 7., 8., 2., 6., 9., 0.],
#        [7., 8., 6., 1., 1., 1., 9., 8.]])]


'复制操作1：深复制，相当于同一对象'
a = np.arange(12)
b = a
print(b is a)
# True
b. shape = (3, 4)
print(a.shape)
# (3, 4)
print(id(a))
# 2007509792608
print(id(b))
# 2007509792608
'复制操作2：浅复制，不同对象，但是操作的值是统一的'
c = a.view()
print(c is a)
# False
print(a)
# [[ 0  1  2  3]
#  [ 4  5  6  7]
#  [ 8  9 10 11]]
c[0, 0] = 123
print(a)
# [[123   1   2   3]
#  [  4   5   6   7]
#  [  8   9  10  11]]
print(id(a))
# 2827766237024
print(id(b))
# 2827766237024
'复制操作3：直接复制，不同对象，互不关联'
d = a.copy()
print(d is a)
# False
print(a)
# [[123   1   2   3]
#  [  4   5   6   7]
#  [  8   9  10  11]]
d[0, 0] = 999
print(d)
# [[999   1   2   3]
#  [  4   5   6   7]
#  [  8   9  10  11]]
print(a)
# [[123   1   2   3]
#  [  4   5   6   7]
#  [  8   9  10  11]]
print(id(a))
# 2260966013008
print(id(d))
# 2260816938688


data = np.sin(np.arange(20)).reshape(5, 4)
print(data)
# [[ 0.          0.84147098  0.90929743  0.14112001]
#  [-0.7568025  -0.95892427 -0.2794155   0.6569866 ]
#  [ 0.98935825  0.41211849 -0.54402111 -0.99999021]
#  [-0.53657292  0.42016704  0.99060736  0.65028784]
#  [-0.28790332 -0.96139749 -0.75098725  0.14987721]]
'找到每列最大值的下标，axis=0表示按列'
ind = data.argmax(axis=0)
print(ind)
# [2 0 3 1]
print(range(data.shape[1]))
# range(0, 4)
data_max = data[ind, range(data.shape[1])]
print(data_max)
# [0.98935825 0.84147098 0.99060736 0.6569866 ]

a = np.arange(0, 40, 10)
print(a)
# [ 0 10 20 30]
'对矩阵整体变换'
b = np.tile(a, (3, 5))
print(b)
# [[ 0 10 20 30  0 10 20 30  0 10 20 30  0 10 20 30  0 10 20 30]
#  [ 0 10 20 30  0 10 20 30  0 10 20 30  0 10 20 30  0 10 20 30]
#  [ 0 10 20 30  0 10 20 30  0 10 20 30  0 10 20 30  0 10 20 30]]


a = np.array([
    [4, 3, 5],
    [1, 2, 1]
])
print(a)
# [[4 3 5]
#  [1 2 1]]
b = np.sort(a, axis=1)
print(b)
# [[3 4 5]
#  [1 1 2]]
a.sort(axis=1)
print(a)
# [[3 4 5]
#  [1 1 2]]


a = np.array([4, 3, 1, 2])
'得出升序后下标'
j = np.argsort(a)
print(j)
# [2 3 1 0]
print(a[j])
# [1 2 3 4]

IO操作

四种存取方式

import numpy as np

'采用二进制存取单个矩阵数组'
a = np.array([1, 2, 3, 4, 5])
'保存到 outfile.npy 文件上'
np.save('outfile.npy', a)
'读取文件'
b = np.load('outfile.npy')
print(b)
# [1 2 3 4 5]


'采用二进制存取多个矩阵数组'
a = np.array([
    [1, 2, 3],
    [4, 5, 6]
])
b = np.arange(0, 1.0, 0.1)
c = np.sin(b)
np.savez("runoob.npz", a=a, b=b, sin_array=c)

r = np.load("runoob.npz")
'查看各个数组名称'
print(r.files)
# ['a', 'b', 'sin_array']
'数组 a'
print(r["a"])
# [[1 2 3]
#  [4 5 6]]
'数组 b'
print(r["b"])
# [0.  0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9]
'数组 c'
print(r["sin_array"])
# [0.         0.09983342 0.19866933 0.29552021 0.38941834 0.47942554
#  0.56464247 0.64421769 0.71735609 0.78332691]


'采用文本存取矩阵数组'
a = np.array([1, 2, 3, 4, 5])
np.savetxt('out.txt', a)

b = np.loadtxt('out.txt')
print(b)
# [1. 2. 3. 4. 5.]


'以二进制的制定格式存取矩阵数组'
a = np.arange(0, 10, 0.5).reshape(4, -1)
'改为保存为整数，以逗号分隔'
np.savetxt("out.txt", a, fmt="%d", delimiter=",")
'load 时也要指定为逗号分隔'
b = np.loadtxt("out.txt", delimiter=",")
print(b)