numpy

标签：argmax   power   split   对象   笔记   字节   art   改变   随机数   

import numpy as np

np.arange(10)

array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

# 对列表中的元素开平方
b = [3, 4, 9]
np.sqrt(b)

array([1.73205081, 2.        , 3.        ])

# 使用array函数创建一维数组
a = np.array([1,2,3,4])
a

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

# 使用array创建二维数组
b = np.array([[1,2,3],[3,4,6]])
b

array([[1, 2, 3],
       [3, 4, 6]])

## 使用dtype参数来设置数组的类型
c = np.array([3,4,5], dtype=float)
c

array([3., 4., 5.])

## 使用ndim参数来设置数组的维度
d = np.array([3,4,6], dtype=float,ndmin=3)
d

array([[[3., 4., 6.]]])

使用arange创建数组

e = np.arange(1,11)
e

array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10])

## 设置步长
np.arange(1,11,2)

array([1, 3, 5, 7, 9])

## 设置dtype
np.arange(1,11,2,dtype=float)

array([1., 3., 5., 7., 9.])

随机数

## random函数创建0.0-1.0之间的随机数数组，注意不包括号1.0
np.random.random(size=5)   ### 一维数组

array([0.14109248, 0.57757689, 0.31587811, 0.90959059, 0.25585209])

## 创建二维数组
np.random.random(size=(3,4)) ### 3,4 表示3行吧列

array([[0.27963513, 0.28511459, 0.63027425, 0.99946311],
       [0.14654823, 0.62364526, 0.81366614, 0.88192038],
       [0.91244454, 0.4028762 , 0.11921253, 0.54991022]])

### 创建三维数组
np.random.random(size=(2,3,4))   ### 2，3，4表示创建2个3行4列

array([[[0.27409587, 0.11804355, 0.34964509, 0.93155881],
        [0.32850319, 0.42947898, 0.21363423, 0.94016219],
        [0.66905712, 0.72171425, 0.23520955, 0.4893854 ]],

       [[0.17257756, 0.99171624, 0.31052962, 0.61989267],
        [0.84792176, 0.79669383, 0.92678657, 0.90099817],
        [0.35242589, 0.95967321, 0.00670096, 0.91882932]]])

随机整数

np.random.randint(6,size=10) # 生成0到5之间的10个随机整数

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

## 生成5-10之间的随机整数的二维数组
np.random.randint(5,11,size=(4,3))

array([[ 9,  7, 10],
       [10,  7,  5],
       [ 8,  8,  7],
       [ 7,  7, 10]])

## 生成三维数组
np.random.randint(5,11,size=(2,4,3))

array([[[ 7,  8,  5],
        [ 7,  5,  6],
        [ 8,  8,  8],
        [ 8,  5, 10]],

       [[ 7, 10,  6],
        [ 6,  8,  8],
        [ 9,  6,  6],
        [ 5,  8,  6]]])

np.random.randint(5,11,size=(1,3,4))

array([[[ 7, 10,  5, 10],
        [ 9, 10,  5, 10],
        [ 6,  9, 10,  9]]])

randint中参数dtype的使用

d = np.random.randint(10,size=5)

d.dtype

dtype(‘int32‘)

f = np.random.randint(10,size=5,dtype=np.int64)

f.dtype

dtype(‘int64‘)

正态分布

## 一维
np.random.randn(4)

array([-0.65738946, -1.36270863, -0.5901353 ,  0.63707697])

### 二维
np.random.randn(2,3)

array([[-0.86151919, -1.35400762,  0.60027059],
       [-0.44439982, -0.69328396,  1.15670548]])

### 三维
np.random.randn(2,3,4)

array([[[ 0.21530965,  1.419034  ,  1.2442361 , -1.47677229],
        [ 0.41533172, -0.8700582 ,  1.28418702,  0.77605165],
        [-1.63712476, -0.24687974,  1.38743355, -0.13334918]],

       [[-0.70745225, -0.73391537, -1.29003062, -0.96750294],
        [-0.48278997, -0.3381754 ,  2.30733888, -1.00871624],
        [-0.88437762,  0.68254093,  0.90741469, -0.70638608]]])

以上是标准的正态分布：期望为0，方差为1

创建指定期望和方差值的正态分布

np.random.normal(size=5) ### 默认的期望是loc=0.0,方差是scale=1.0

array([ 0.37960069,  0.77483781,  0.23537764,  0.97151935, -0.08167272])

### 指定loc和scale
np.random.normal(loc=2,scale=3,size=5)

array([ 5.49844885,  6.50644569, -2.05466087,  2.53766272,  2.40695409])

np.random.normal(loc=2,scale=3,size=(3,4))

array([[ 6.95415439,  2.37018031, -2.9699421 , -1.12177319],
       [ 2.69676034,  0.30211155,  4.04696743, -1.63856255],
       [ 2.90005922, -3.83970821,  3.07125062,  3.78305087]])

ndarray对象属性

### 分别创建一维，二维，三维数组
a1 = np.array([1,2,3,4,5])
b1 = np.random.randint(4,10,size=(3,4))
c1 = np.random.randn(2,3,4)

a1

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

b1

array([[6, 9, 7, 6],
       [7, 5, 7, 5],
       [9, 8, 7, 7]])

c1

array([[[-1.70317407,  0.57392171, -0.00487497, -0.32478167],
        [ 0.69572239, -0.34328439, -1.31930338, -0.26807493],
        [ 1.22140779, -0.49691301,  0.69210554,  0.14662866]],

       [[ 0.083173  , -1.14836617,  1.5909447 ,  0.55987063],
        [-0.59362023, -0.73479667,  0.78516186, -0.46653616],
        [-0.02626701, -1.49199613,  1.2136789 ,  2.01738911]]])

ndim属性

print(a1.ndim,b1.ndim,c1.ndim)

1 2 3

shape属性

print(a1.shape,b1.shape,c1.shape)

(5,) (3, 4) (2, 3, 4)

dtype属性

print(a1.dtype,b1.dtype,c1.dtype)

int32 int32 float64

size属性，元素的总个数

print(a1.size,b1.size,c1.size)

5 12 24

itemsize 每个元素所占的字节

print(a1.itemsize,b1.itemsize,c1.itemsize)

4 4 8

zeros创建数组

np.zeros(5)

array([0., 0., 0., 0., 0.])

np.zeros(5,dtype=int)

array([0, 0, 0, 0, 0])

np.zeros((3,4))

array([[0., 0., 0., 0.],
       [0., 0., 0., 0.],
       [0., 0., 0., 0.]])

ones创建数组

np.ones(10)

array([1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

np.ones((2,5),dtype=int)

array([[1, 1, 1, 1, 1],
       [1, 1, 1, 1, 1]])

empty

np.empty(8)

array([0.0000000e+000, 0.0000000e+000, 0.0000000e+000, 0.0000000e+000,
       0.0000000e+000, 7.3516968e-321, 1.3796137e-306, 0.0000000e+000])

np.empty((3,4))

array([[0., 0., 0., 0.],
       [0., 0., 0., 0.],
       [0., 0., 0., 0.]])

linspace 创建等差数列

np.linspace(1,10,10)

array([ 1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9., 10.])

np.linspace(5,20,5)

array([ 5.  ,  8.75, 12.5 , 16.25, 20.  ])

np.linspace(5,20,5,endpoint=False) #### endpoint默认为True

array([ 5.,  8., 11., 14., 17.])

logspace创建等比数列

np.logspace(0,9,10,base=2) ### base默认为空0。0

array([  1.,   2.,   4.,   8.,  16.,  32.,  64., 128., 256., 512.])

numpy 索引与切片

d1 = np.arange(10)

d1

array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

d1[0]

0

d1[-1]

9

### 切片[start:stop:step]
d1[:]

array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

d1[3:]

array([3, 4, 5, 6, 7, 8, 9])

d1[3:5]

array([3, 4])

d1[1:7:2]

array([1, 3, 5])

d1[::-1]  # -1表示反向获取

array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0])

d1[-5:-2]

array([5, 6, 7])

d1[-7:-2:2]

array([3, 5, 7])

二维数组的切片

x = np.arange(1,13)

x

array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12])

### 对一维数组修改形状
x1 = x.reshape(4,3)

x1

array([[ 1,  2,  3],
       [ 4,  5,  6],
       [ 7,  8,  9],
       [10, 11, 12]])

x1[1]

array([4, 5, 6])

x1[1][2]

6

### 二维数组切片【行进行切片，列进行切片】
x1[:,:]

array([[ 1,  2,  3],
       [ 4,  5,  6],
       [ 7,  8,  9],
       [10, 11, 12]])

x1[:,1]

array([ 2,  5,  8, 11])

x1[:,0:2]  ### 所有行第一列第二列

array([[ 1,  2],
       [ 4,  5],
       [ 7,  8],
       [10, 11]])

x1[::2,:]   #### 奇数行所有列

array([[1, 2, 3],
       [7, 8, 9]])

x1[::2,0:2]  ## 奇数行，第一，二列

array([[1, 2],
       [7, 8]])

坐标获取

x1[1,2] 

6

x1[(1,2),(2,0)]

array([6, 7])

x1[-1]

array([10, 11, 12])

x1[::-1]

array([[10, 11, 12],
       [ 7,  8,  9],
       [ 4,  5,  6],
       [ 1,  2,  3]])

x1[::-1,::-1]

array([[12, 11, 10],
       [ 9,  8,  7],
       [ 6,  5,  4],
       [ 3,  2,  1]])

数组的复制

sub_a = x1[:2,:2]

sub_a

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

sub_a[0][0] =100

sub_a

array([[100,   2],
       [  4,   5]])

x1

array([[100,   2,   3],
       [  4,   5,   6],
       [  7,   8,   9],
       [ 10,  11,  12]])

### 通过切片可以获取到新的数组，即使赋值给新的变量，但是还是原来数组的视图，如果对切片数组中元素的值进行修改，原数组中也会改变
### numpy中的copy实现了深 考贝
sub = np.copy(x1[:2,:2])

sub

array([[100,   2],
       [  4,   5]])

sub[0][1] = 21

sub

array([[100,  21],
       [  4,   5]])

x1

array([[100,   2,   3],
       [  4,   5,   6],
       [  7,   8,   9],
       [ 10,  11,  12]])

改变数组的维度

w = np.arange(1,25)

w

array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
       18, 19, 20, 21, 22, 23, 24])

w.reshape(4,6)  ###二维

array([[ 1,  2,  3,  4,  5,  6],
       [ 7,  8,  9, 10, 11, 12],
       [13, 14, 15, 16, 17, 18],
       [19, 20, 21, 22, 23, 24]])

w.reshape((3,8))

array([[ 1,  2,  3,  4,  5,  6,  7,  8],
       [ 9, 10, 11, 12, 13, 14, 15, 16],
       [17, 18, 19, 20, 21, 22, 23, 24]])

### 三维
w.reshape((2,3,4))

array([[[ 1,  2,  3,  4],
        [ 5,  6,  7,  8],
        [ 9, 10, 11, 12]],

       [[13, 14, 15, 16],
        [17, 18, 19, 20],
        [21, 22, 23, 24]]])

np.reshape(w,(3,8))

array([[ 1,  2,  3,  4,  5,  6,  7,  8],
       [ 9, 10, 11, 12, 13, 14, 15, 16],
       [17, 18, 19, 20, 21, 22, 23, 24]])

np.reshape(w,(4,3,2))

array([[[ 1,  2],
        [ 3,  4],
        [ 5,  6]],

       [[ 7,  8],
        [ 9, 10],
        [11, 12]],

       [[13, 14],
        [15, 16],
        [17, 18]],

       [[19, 20],
        [21, 22],
        [23, 24]]])

将多维修改成一维数组

bb = np.reshape(w,(2,4,3))

bb

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9],
        [10, 11, 12]],

       [[13, 14, 15],
        [16, 17, 18],
        [19, 20, 21],
        [22, 23, 24]]])

bb.reshape(24)

array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
       18, 19, 20, 21, 22, 23, 24])

bb.reshape(-1)

array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
       18, 19, 20, 21, 22, 23, 24])

bb.ravel()

array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
       18, 19, 20, 21, 22, 23, 24])

bb.flatten()

array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
       18, 19, 20, 21, 22, 23, 24])

数组的拼接

### 创建两个二维数组
h1 = np.array([[1,2,3],[4,5,6]])
h2 = np.array([[11,12,13],[14,15,16]])

h1

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

h2

array([[11, 12, 13],
       [14, 15, 16]])

### 小平拼接
np.hstack([h1,h2])

array([[ 1,  2,  3, 11, 12, 13],
       [ 4,  5,  6, 14, 15, 16]])

np.hstack((h1,h2))

array([[ 1,  2,  3, 11, 12, 13],
       [ 4,  5,  6, 14, 15, 16]])

#### 垂直拼接
np.vstack((h1,h2))

array([[ 1,  2,  3],
       [ 4,  5,  6],
       [11, 12, 13],
       [14, 15, 16]])

### concatenate拼接
np.concatenate((h1,h2),axis=0)   ## axis=0是默认方向，垂直方向

array([[ 1,  2,  3],
       [ 4,  5,  6],
       [11, 12, 13],
       [14, 15, 16]])

np.concatenate((h1,h2))

array([[ 1,  2,  3],
       [ 4,  5,  6],
       [11, 12, 13],
       [14, 15, 16]])

np.concatenate((h1,h2),axis=1)

array([[ 1,  2,  3, 11, 12, 13],
       [ 4,  5,  6, 14, 15, 16]])

### 三维数组 axis有0,1,2
one = np.arange(1,13).reshape(1,2,6)
two = np.arange(101,113).reshape(1,2,6)

one

array([[[ 1,  2,  3,  4,  5,  6],
        [ 7,  8,  9, 10, 11, 12]]])

two

array([[[101, 102, 103, 104, 105, 106],
        [107, 108, 109, 110, 111, 112]]])

np.concatenate((one,two),axis=0)

array([[[  1,   2,   3,   4,   5,   6],
        [  7,   8,   9,  10,  11,  12]],

       [[101, 102, 103, 104, 105, 106],
        [107, 108, 109, 110, 111, 112]]])

np.concatenate((one,two),axis=1)

array([[[  1,   2,   3,   4,   5,   6],
        [  7,   8,   9,  10,  11,  12],
        [101, 102, 103, 104, 105, 106],
        [107, 108, 109, 110, 111, 112]]])

np.concatenate((one,two),axis=2)

array([[[  1,   2,   3,   4,   5,   6, 101, 102, 103, 104, 105, 106],
        [  7,   8,   9,  10,  11,  12, 107, 108, 109, 110, 111, 112]]])

数组的分隔

s = np.arange(1,13)

s

array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12])

### 平均分隔
np.split(s,4)

[array([1, 2, 3]), array([4, 5, 6]), array([7, 8, 9]), array([10, 11, 12])]

### 按位置分隔
np.split(s,(4,6))

[array([1, 2, 3, 4]), array([5, 6]), array([ 7,  8,  9, 10, 11, 12])]

#### 二维
ss = np.array([[1,2,3,4],[5,6,7,8],[8,10,11,12],[13,14,15,16]])

ss

array([[ 1,  2,  3,  4],
       [ 5,  6,  7,  8],
       [ 8, 10, 11, 12],
       [13, 14, 15, 16]])

np.split(ss,2,axis=0)

[array([[1, 2, 3, 4],
        [5, 6, 7, 8]]), array([[ 8, 10, 11, 12],
        [13, 14, 15, 16]])]

np.split(ss,(2,3),axis=0)

[array([[1, 2, 3, 4],
        [5, 6, 7, 8]]), array([[ 8, 10, 11, 12]]), array([[13, 14, 15, 16]])]

np.split(ss,2,axis=1)

[array([[ 1,  2],
        [ 5,  6],
        [ 8, 10],
        [13, 14]]), array([[ 3,  4],
        [ 7,  8],
        [11, 12],
        [15, 16]])]

np.split(ss,[3],axis=1)

[array([[ 1,  2,  3],
        [ 5,  6,  7],
        [ 8, 10, 11],
        [13, 14, 15]]), array([[ 4],
        [ 8],
        [12],
        [16]])]

## hsplit()小平方向分隔
np.hsplit(ss,2)

[array([[ 1,  2],
        [ 5,  6],
        [ 8, 10],
        [13, 14]]), array([[ 3,  4],
        [ 7,  8],
        [11, 12],
        [15, 16]])]

np.hsplit(ss,[3])

[array([[ 1,  2,  3],
        [ 5,  6,  7],
        [ 8, 10, 11],
        [13, 14, 15]]), array([[ 4],
        [ 8],
        [12],
        [16]])]

np.vsplit(ss,2)  ### vsplit()垂直分隔

[array([[1, 2, 3, 4],
        [5, 6, 7, 8]]), array([[ 8, 10, 11, 12],
        [13, 14, 15, 16]])]

np.vsplit(ss,[1])

[array([[1, 2, 3, 4]]), array([[ 5,  6,  7,  8],
        [ 8, 10, 11, 12],
        [13, 14, 15, 16]])]

数组的转置

### 创建二维数组
tr = np.arange(1,25).reshape(8,3)

tr

array([[ 1,  2,  3],
       [ 4,  5,  6],
       [ 7,  8,  9],
       [10, 11, 12],
       [13, 14, 15],
       [16, 17, 18],
       [19, 20, 21],
       [22, 23, 24]])

## 使用transpose a[j][i]  --->a[i][j]
tr.transpose()

array([[ 1,  4,  7, 10, 13, 16, 19, 22],
       [ 2,  5,  8, 11, 14, 17, 20, 23],
       [ 3,  6,  9, 12, 15, 18, 21, 24]])

tr.T

array([[ 1,  4,  7, 10, 13, 16, 19, 22],
       [ 2,  5,  8, 11, 14, 17, 20, 23],
       [ 3,  6,  9, 12, 15, 18, 21, 24]])

np.transpose(tr)

array([[ 1,  4,  7, 10, 13, 16, 19, 22],
       [ 2,  5,  8, 11, 14, 17, 20, 23],
       [ 3,  6,  9, 12, 15, 18, 21, 24]])

## 多维数组
trs = tr.reshape(2,3,4)

trs

array([[[ 1,  2,  3,  4],
        [ 5,  6,  7,  8],
        [ 9, 10, 11, 12]],

       [[13, 14, 15, 16],
        [17, 18, 19, 20],
        [21, 22, 23, 24]]])

np.transpose(trs)  ## 默认是从a[i][j][k]  ---->a[k][j][i]

array([[[ 1, 13],
        [ 5, 17],
        [ 9, 21]],

       [[ 2, 14],
        [ 6, 18],
        [10, 22]],

       [[ 3, 15],
        [ 7, 19],
        [11, 23]],

       [[ 4, 16],
        [ 8, 20],
        [12, 24]]])

np.transpose(trs,(1,2,0)) ###  这里的1，2，0是数组维度的下标

array([[[ 1, 13],
        [ 2, 14],
        [ 3, 15],
        [ 4, 16]],

       [[ 5, 17],
        [ 6, 18],
        [ 7, 19],
        [ 8, 20]],

       [[ 9, 21],
        [10, 22],
        [11, 23],
        [12, 24]]])

函数

###创建一个二维数组
v2 = np.arange(9).reshape(3,3)

v2

array([[0, 1, 2],
       [3, 4, 5],
       [6, 7, 8]])

### 创建一个一维数组
v1 = np.array([10,10,10])

v1

array([10, 10, 10])

### 算术函数
np.add(v1,v2)

array([[10, 11, 12],
       [13, 14, 15],
       [16, 17, 18]])

v1+v2

array([[10, 11, 12],
       [13, 14, 15],
       [16, 17, 18]])

np.subtract(v1,v2)

array([[10,  9,  8],
       [ 7,  6,  5],
       [ 4,  3,  2]])

### out参数的使用
y = np.empty((3,3),dtype=np.int)
np.multiply(v2,10,out=y)

array([[ 0, 10, 20],
       [30, 40, 50],
       [60, 70, 80]])

### 三角函数
np.sin(np.array([0,30,60,90]))

array([ 0.        , -0.98803162, -0.30481062,  0.89399666])

### 向上取整
f = np.array([1.0,4.55,123,0.567,25.33])
np.around(f)

array([  1.,   5., 123.,   1.,  25.])

np.ceil(f)

array([  1.,   5., 123.,   1.,  26.])

## 向下取整
np.floor(f)

array([  1.,   4., 123.,   0.,  25.])

统计函数

### power()幂次方
t = np.arange(1,13).reshape(3,4)

t

array([[ 1,  2,  3,  4],
       [ 5,  6,  7,  8],
       [ 9, 10, 11, 12]])

np.power(t,2)

array([[  1,   4,   9,  16],
       [ 25,  36,  49,  64],
       [ 81, 100, 121, 144]], dtype=int32)

## power（）中参数out的使用
x = np.arange(5)
y = np.zeros(10)

np.power(2,x,out=y[:5])

array([ 1.,  2.,  4.,  8., 16.])

y

array([ 1.,  2.,  4.,  8., 16.,  0.,  0.,  0.,  0.,  0.])

求中位数

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

np.median(z) ### 对数组排序 [1,2,2,3,4,5]  数组中元素个数为偶数 中位数指：中间两个数的平均值

2.5

np.median(np.array([4,3,2,5,6])) # 对数组排序 [2,2,3,4,5]  数组中元素个数为奇数 中位数指：中间的数

4.0

### 二维数组
z2 = np.arange(1,13).reshape(3,4)

z2

array([[ 1,  2,  3,  4],
       [ 5,  6,  7,  8],
       [ 9, 10, 11, 12]])

np.median(z2,axis=0)

array([5., 6., 7., 8.])

np.median(z2,axis=1)

array([ 2.5,  6.5, 10.5])

mean求平均值

np.mean(z)

2.8333333333333335

np.mean(z2)

6.5

np.mean(z2,axis=0)

array([5., 6., 7., 8.])

np.mean(z2,axis=1)

array([ 2.5,  6.5, 10.5])

#sum()
np.max(z)

5

np.sum(z)

17

np.min(z)

1

np.argmax(z) ### 返回最大值为的下标

3

np.argmin(z)

5

numpy

标签：argmax   power   split   对象   笔记   字节   art   改变   随机数   

原文地址：https://www.cnblogs.com/baiyifengyun/p/14101768.html