numpy

Logout numpy Last Checkpoint: 06/11/2018 (autosaved)

Python 3 

Not Trusted

• File
• Edit
• View
• Insert
• Cell
• Kernel
• Widgets
• Help

CodeMarkdownRaw NBConvertHeading-

导入numpy并查看版本

In [2]:

import numpy as np 

​

np.__version__

​

Out[2]:

'1.13.1'

什么是numpy？

Numpy即Numeric Python，python经过扩展可以支持数组和矩阵类型，并且具有大量的函数可以计算这些数组和矩阵。这些数组一般是多维的，而这个扩展的程序包就是numpy。

【注】是数据分析和机器学习中最基本的工具，后面许多API和工具都是建立在他得基础上，如：pandas、scipy、matplotlib等

一、创建ndarray

numpy中最基础数据结构就是ndarray：即数组

1. 使用np.array()由python list创建

In [3]:

data = [1,2,3]

nd = np.array(data)

nd

Out[3]:

array([1, 2, 3])

In [3]:

data

Out[3]:

[1, 2, 3]

In [4]:

type(nd),type(data)

Out[4]:

(numpy.ndarray, list)

In [5]:

nd.dtype # 数组中的元素的类型

Out[5]:

dtype('int32')

In [8]:

nd2 = np.array([1,2,3,4.5,True,"1234"])

nd2

Out[8]:

array(['1', '2', '3', '4.5', 'True', '1234'],
      dtype='<U32')

In [7]:

nd2.dtype

Out[7]:

dtype('float64')

注意：1、数组中所有的元素类型相同 2、如果通过列表来创建的时候，列表中元素不一样，会被统一成某种类型（优先级：str>float>int）

In [9]:

# 注意：图片也可在numpy里面用数组来表示

# 引入一张图片

import matplotlib.pyplot as plt # 这个是用于处理数据可视化的一个工具

In [165]:

girl = plt.imread("./source/girl.jpg")

girl # 彩色图片是一个三维的数组

Out[165]:

array([[[225, 231, 231],
        [229, 235, 235],
        [222, 228, 228],
        ..., 
        [206, 213, 162],
        [211, 213, 166],
        [217, 220, 173]],

       [[224, 230, 230],
        [229, 235, 235],
        [223, 229, 229],
        ..., 
        [206, 213, 162],
        [211, 213, 166],
        [217, 220, 173]],

       [[224, 230, 230],
        [229, 235, 235],
        [223, 229, 229],
        ..., 
        [206, 213, 162],
        [211, 213, 166],
        [219, 221, 174]],

       ..., 
       [[175, 187, 213],
        [180, 192, 218],
        [175, 187, 213],
        ..., 
        [155, 162, 180],
        [153, 160, 178],
        [156, 163, 181]],

       [[175, 187, 213],
        [180, 192, 218],
        [174, 186, 212],
        ..., 
        [155, 162, 180],
        [153, 160, 178],
        [155, 162, 180]],

       [[177, 189, 215],
        [181, 193, 219],
        [174, 186, 212],
        ..., 
        [155, 162, 180],
        [153, 160, 178],
        [156, 163, 181]]], dtype=uint8)

In [11]:

girl.shape # 数据的形状

Out[11]:

(900, 1440, 3)

In [13]:

plt.imshow(girl)

plt.show()

In [14]:

boy = np.array([[0.21,0.14,0.12343,0.14],

                [0.21,0.14,0.12343,0.14],

                [0.21,0.14,0.12343,0.14],

               ])

In [17]:

plt.imshow(boy,cmap="gray")

plt.show()

In [19]:

g = girl[:200,:300]

In [20]:

plt.imshow(g)

plt.show()

In [ ]:

​

2. 使用np的routines函数创建

1)np.ones(shape,dtype=None,order='C')

In [26]:

# 参数shape数据的形状传一个元组，元组的第0个元素代表第0维中的元素个数，依次类推

np.ones((3,2,4,5,6,7,8))

In [28]:

ones

Out[28]:

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

       [[ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        ..., 
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        ..., 
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.]],

       ..., 
       [[ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        ..., 
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        ..., 
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        ..., 
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.],
        [ 1.,  1.,  1.]]])

In [27]:

ones = np.ones((168,59,3))

plt.imshow(ones)

plt.show()

In [29]:

ones[:,:,1:] = 0 # 数组是可以切片赋值的

In [34]:

ones[:,:,0] = 0.3

In [35]:

ones

Out[35]:

array([[[ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        ..., 
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ]],

       [[ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        ..., 
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ]],

       [[ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        ..., 
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ]],

       ..., 
       [[ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        ..., 
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ]],

       [[ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        ..., 
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ]],

       [[ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        ..., 
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ],
        [ 0.3,  0. ,  0. ]]])

In [36]:

plt.imshow(ones)

plt.show()

In [ ]:

​

2）np.zeros(shape,dtype="float",order="C")

In [40]:

np.zeros((2,3),dtype="int32")

Out[40]:

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

3）np.full(shape,fill_value,dtype=None)

In [43]:

np.full((3,4),12,dtype="float")

Out[43]:

array([[ 12.,  12.,  12.,  12.],
       [ 12.,  12.,  12.,  12.],
       [ 12.,  12.,  12.,  12.]])

4）np.eye(N,M,k=0,dtype='float')

In [44]:

np.eye(4)

Out[44]:

array([[ 1.,  0.,  0.,  0.],
       [ 0.,  1.,  0.,  0.],
       [ 0.,  0.,  1.,  0.],
       [ 0.,  0.,  0.,  1.]])

5）np.linspace(start,stop,num=50)

In [49]:

np.linspace(0,10,10)

Out[49]:

array([  0.        ,   1.11111111,   2.22222222,   3.33333333,
         4.44444444,   5.55555556,   6.66666667,   7.77777778,
         8.88888889,  10.        ])

In [50]:

np.logspace(1,100,10)

Out[50]:

array([  1.00000000e+001,   1.00000000e+012,   1.00000000e+023,
         1.00000000e+034,   1.00000000e+045,   1.00000000e+056,
         1.00000000e+067,   1.00000000e+078,   1.00000000e+089,
         1.00000000e+100])

In [ ]:

​

6）np.arange([start,]stop,[step,]dtype=None) "[]"中是可选项

In [54]:

np.arange(0,100,5)

Out[54]:

array([ 0,  5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80,
       85, 90, 95])

7）np.random.randint(low,high=None,size=None,dtype='I')

In [67]:

np.random.randint(0,10,size=(10,5)) # low下限 high上限 size形状（元组）

Out[67]:

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

8）np.random.randn(d0,d1,...,dn) 从第一维度到第n维度生成一个数组，数组中的数字符合标准正态分布

In [70]:

np.random.randn(5,6,10) # N(0,1)

Out[70]:

array([[[-0.09358215, -1.49899141,  1.98769631,  1.51193985, -0.10427579,
         -0.39974708, -0.0359005 ,  0.73853509,  1.57513744,  0.2359057 ],
        [-0.7041393 ,  0.5482207 , -1.47905978, -1.86071464,  0.24794821,
         -1.85676   ,  0.10820599, -0.32210508,  0.22393499, -0.66522583],
        [ 1.60692141,  1.08030533, -0.82684758,  0.11892321,  1.60067252,
          0.78420728, -1.02287912,  0.13980319, -0.35365909,  1.89048322],
        [-0.79423431, -1.13066957,  0.29044997,  1.83286616,  1.01731098,
          0.36688754, -0.78547307, -0.03262272, -1.63958452,  0.52841964],
        [ 0.63854724, -0.83270504,  2.33109623, -2.0670137 , -1.10680899,
          0.06584844,  0.39706437,  0.55065026, -0.59335951,  0.22210873],
        [ 0.41064818,  1.68676676,  0.30244373,  0.0282829 ,  0.19600649,
         -0.68277526,  0.06810738,  0.53271374, -0.4204973 ,  0.71823602]],

       [[-0.32284878, -0.76235291,  0.98673175,  1.91000432,  0.76937097,
          0.16359419, -1.17058448,  0.73993218,  1.29904232, -1.41618621],
        [-0.37633528,  0.07615863,  0.37582282, -1.63055539,  1.37079158,
         -0.61601884,  0.39195051,  1.39733195, -0.60085489,  0.04125237],
        [ 0.26121747,  0.4540636 , -0.94938301, -1.44742311, -0.46626518,
          0.47067078, -1.49014315, -0.10415939,  2.05483157,  0.25061365],
        [-0.67645451, -0.78386469,  0.9837996 ,  1.10329857, -0.63704624,
          1.36715299,  1.3569239 , -0.72635522,  0.51392834,  0.82614783],
        [-0.23384279, -1.28411627, -0.05939828,  0.44944251,  1.189167  ,
          0.87113199,  0.53838983,  0.41908957, -0.74518935,  1.16339158],
        [-0.18668839, -0.4712675 , -0.06185288,  0.8182108 , -1.18638766,
         -1.73182272, -0.19059271, -0.11019542, -0.45788589,  1.6559319 ]],

       [[ 0.13107506,  0.09003381, -0.04046429, -0.87251826, -2.50623334,
         -0.06766936,  0.87597133, -1.53979928, -0.93540774,  0.45923245],
        [ 0.47169492,  0.16740181,  0.46952713,  1.19222423, -1.3851465 ,
          1.99008688,  1.50716764, -0.89799582, -0.05838133,  0.40384442],
        [-0.54836846, -0.42615348, -0.5839804 ,  0.03183017, -0.18666256,
          0.26916727, -0.20821308,  0.65122173, -0.21675408, -0.93981481],
        [ 0.39832431, -1.3087661 , -0.58729167,  1.12946505,  0.32596195,
         -0.07288169, -0.60721266,  2.33748491, -1.43005908, -0.99483587],
        [ 1.08070407,  1.06087286,  0.53105574,  0.15470968,  0.56741346,
         -0.43695276, -0.51750765,  1.13183184,  0.49298563, -0.87541634],
        [-2.06517303, -0.0877284 , -1.22371533, -1.59688986, -0.4547517 ,
         -0.40808438, -0.35837045, -1.54584721, -0.19055862, -1.0929987 ]],

       [[ 0.75912634,  0.67984287,  0.86651616,  1.44612227, -0.50480375,
         -0.86431858, -1.49997546,  0.60060329, -0.76905209, -0.1248285 ],
        [-0.08862881,  0.79658324, -0.54596008, -0.47540929, -1.45907461,
          0.07354412, -1.26479651,  1.40530143,  0.60891271,  1.07599106],
        [ 0.44097311,  0.10543125, -1.64720857, -0.49178975, -0.40010248,
          0.60104546,  0.61886559,  0.09964341, -0.07654484,  0.01466248],
        [ 0.38690965, -0.64976725, -0.88243344,  0.43953291,  0.42834801,
          0.23491374, -0.18973705,  1.27917483,  0.82983269,  0.79430021],
        [ 0.91834613,  1.12247777,  0.08971992, -0.45914651,  0.79840117,
          1.23251681, -0.12130413, -0.13878377, -0.09306616,  0.43146146],
        [-0.86534235,  1.20442112,  0.80460288,  0.23146194, -1.37131063,
          0.39225933, -2.64398314, -0.37121153, -0.98828702, -0.48872689]],

       [[-0.95197633,  0.40663767,  0.30596473, -1.13367237,  0.56080945,
         -0.18929488, -1.00550724,  1.82465218,  0.97053687, -0.18880776],
        [ 0.56076673,  0.45680001,  0.04403547, -0.33633002,  1.05787698,
         -0.95026294,  1.55784308, -0.04700079, -1.13309294, -1.52973685],
        [-1.05473426, -0.35119816, -0.41996786, -1.41545262,  1.37013952,
          0.44812898,  0.44133247, -0.69188934, -1.6819927 ,  0.03398214],
        [ 0.2265335 , -1.92836525, -0.31944456, -0.37962678, -0.58608731,
          1.32122153,  1.55326426,  0.49838209,  0.45043623, -1.54473891],
        [ 0.97935231, -0.42807334, -1.29984072,  0.23391228, -1.18660991,
         -0.19844072, -1.83693156, -0.87926028, -1.41108227, -1.66544272],
        [ 1.51680404, -0.90072061, -1.30384261,  1.46910357,  1.55454131,
         -0.14601832,  0.15141248, -1.87883605,  0.0924951 ,  1.28106374]]])

9）np.random.normal(loc=0.0,scale=1.0,size=None)

In [73]:

np.random.normal(450,100,size=(90,10))

Out[73]:

array([[ 336.75387404,  391.65683168,  498.76644615,  386.17886947,
         327.48245651,  527.80674828,  539.26940105,  369.64475214,
         415.14180776,  387.8046801 ],
       [ 436.64172537,  512.42693445,  509.03160498,  428.37233551,
         384.03217165,  450.93849607,  490.15376541,  603.00894225,
         494.53309375,  481.88682217],
       [ 385.30889406,  476.28508918,  490.30271889,  473.79751503,
         567.20252837,  397.00172353,  412.19518771,  470.26351719,
         585.94598243,  406.30114867],
       [ 662.51560347,  455.63555885,  335.75668217,  553.51109653,
         527.50580761,  418.68255566,  433.27222119,  355.5035559 ,
         552.79303757,  365.12473145],
       [ 545.41809506,  542.81120741,  418.08540191,  502.08035002,
         489.61540484,  603.48703621,  442.5173826 ,  310.62586184,
         397.59667339,  407.18629259],
       [ 536.62705543,  373.17087601,  387.53239657,  493.50100186,
         434.69796839,  390.03931656,  443.45656768,  388.89231904,
         424.15132003,  540.91910513],
       [ 440.52306579,  391.71451744,  437.55706601,  469.27630621,
         406.99478958,  362.44118414,  536.48731847,  572.90132295,
         481.58805791,  478.47964791],
       [ 385.14707894,  469.12943309,  305.70922198,  460.74707662,
         502.95838503,  431.71240517,  342.7916848 ,  407.50858253,
         290.838901  ,  527.05185043],
       [ 385.05812408,  441.5621423 ,  479.94238267,  475.70258941,
         548.79318074,  339.69251585,  457.66895693,  407.6839961 ,
         453.96735266,  355.04312047],
       [ 240.57346971,  462.02119378,  557.47031693,  365.5621482 ,
         421.48731128,  337.17012275,  538.99367394,  291.96240587,
         583.40810146,  392.68739514],
       [ 262.77027625,  389.19618291,  431.91418188,  518.30939231,
         535.60696944,  265.01106923,  205.66965664,  477.7160443 ,
         402.08998341,  461.94465525],
       [ 357.43465369,  439.3195828 ,  253.53609204,  511.6280141 ,
         391.66839943,  381.00739543,  394.91400566,  532.95515557,
         483.49393099,  374.07144475],
       [ 315.81211778,  421.4360574 ,  373.72983086,  382.76578134,
         324.43190939,  426.33506066,  380.28268705,  397.3292164 ,
         329.78748304,  503.23789543],
       [ 285.70801477,  469.69661703,  445.8417405 ,  551.44664511,
         369.90599608,  358.48253945,  402.7896672 ,  573.84596855,
         419.84192647,  460.71713048],
       [ 580.08121766,  450.31393694,  382.84384675,  585.53508672,
         460.57849068,  405.79733799,  171.29825334,  456.17549381,
         525.86256043,  412.1363241 ],
       [ 461.60455293,  151.86396147,  415.79059815,  478.80770885,
         303.21793632,  468.37932735,  448.89482993,  597.60308308,
         398.06184732,  469.54027055],
       [ 633.05994548,  367.54906343,  586.66431898,  532.71822618,
         384.30045719,  471.01613892,  587.02823268,  408.6937843 ,
         490.70406224,  301.23699981],
       [ 557.01114   ,  478.46470521,  359.19499266,  387.36093877,
         392.86815293,  547.70014298,  592.9733081 ,  430.08036393,
         503.47793168,  421.89404986],
       [ 506.10167685,  439.12995299,  494.33281733,  345.60884943,
         462.86859726,  434.69600809,  389.5733517 ,  559.97030485,
         469.61006439,  470.99794124],
       [ 400.74372025,  377.36611363,  440.39145837,  546.22740815,
         446.26322926,  554.6736657 ,  583.22012567,  592.92070545,
         538.93499161,  282.85180835],
       [ 487.93671662,  141.59169637,  439.00168061,  451.57280006,
         545.33727124,  374.18559104,  425.96319208,  535.5891261 ,
         412.10195375,  376.20679553],
       [ 328.29975372,  437.43640405,  267.43331932,  417.88751835,
         623.61080152,  581.53803664,  500.60264026,  357.12209808,
         430.17787952,  541.08416585],
       [ 472.62011773,  456.44480533,  333.70041343,  575.68893734,
         483.15426987,  414.06087041,  442.73896679,  476.77790667,
         557.7860946 ,  602.44875414],
       [ 402.43823668,  367.76123377,  434.69216975,  393.08813671,
         584.32419951,  360.853422  ,  555.40311867,  450.11737225,
         318.99533699,  253.97578605],
       [ 390.31529063,  568.87396775,  551.58155749,  498.13129547,
         656.60712073,  517.22802125,  376.85580205,  296.18382599,
         437.453069  ,  507.81430915],
       [ 472.41212452,  333.78376315,  320.0178681 ,  539.43156964,
         412.89365481,  350.07191578,  472.15131164,  496.32800658,
         362.46820011,  399.043498  ],
       [ 563.27488787,  544.29387737,  364.05194481,  381.92267492,
         458.74888938,  437.37408036,  375.25486806,  435.20877258,
         568.03557044,  407.30568794],
       [ 656.1503502 ,  310.95317562,  495.12074898,  528.93019423,
         286.4628272 ,  327.45044748,  508.11552734,  311.80841338,
         460.5253883 ,  516.5762789 ],
       [ 420.43753082,  386.78902839,  555.86664412,  361.12743763,
         532.04777745,  547.14759241,  457.65703769,  581.46595806,
         345.43303389,  436.91156777],
       [ 383.12409576,  405.42886366,  456.65786045,  430.52458332,
         264.60953026,  386.5731847 ,  515.56284123,  473.60876973,
         483.56760088,  427.29662144],
       [ 299.09917474,  298.90384684,  313.14909255,  485.09782479,
         359.66573231,  528.76712485,  372.6380624 ,  496.60810494,
         331.09851695,  293.35089898],
       [ 562.00358333,  402.43286237,  533.74183027,  297.93570041,
         429.96840678,  458.26365946,  338.38340497,  397.53714394,
         398.22398763,  356.31445468],
       [ 404.77436667,  384.16332639,  517.37154695,  376.1193488 ,
         528.47036871,  310.88788993,  474.88444046,  599.45941831,
         159.65492388,  513.85166262],
       [ 586.79169996,  315.1436137 ,  571.37198315,  401.64412536,
         499.14806795,  353.17259466,  346.85185868,  441.90230762,
         399.66471328,  359.00340789],
       [ 554.53046397,  588.65848933,  158.20879966,  300.23079482,
         477.31516759,  422.05477671,  262.69791109,  226.91873827,
         347.23289159,  509.05077155],
       [ 563.92242222,  379.17608475,  371.22672165,  471.86392331,
         332.70531436,  267.45904342,  466.03347237,  413.37683235,
         639.31512283,  489.67462391],
       [ 422.5460759 ,  423.98173223,  380.79628393,  395.64056863,
         526.35578902,  415.15127071,  292.4837271 ,  357.0199034 ,
         576.95240606,  337.70511632],
       [ 323.6864567 ,  439.03720616,  356.49820378,  379.98997979,
         491.69294389,  452.94410429,  434.60975151,  422.65937394,
         575.07419808,  346.5540995 ],
       [ 363.05583763,  519.22694589,  555.72713564,  521.18823145,
         486.64236301,  362.99434703,  508.12612162,  416.08365947,
         355.96065708,  544.18346451],
       [ 532.99878409,  453.0085492 ,  600.56294399,  396.98757863,
         503.14897901,  247.90875336,  496.14510599,  410.53653883,
         441.41998089,  392.91958738],
       [ 415.66468609,  687.49147825,  475.50418028,  423.85438642,
         578.17001518,  241.02835169,  516.94252816,  572.80451212,
         368.70643657,  435.50322231],
       [ 467.78924383,  409.2163454 ,  391.59584661,  653.29582704,
         417.34844942,  404.70467089,  372.8178365 ,  527.88773237,
         397.43735984,  557.50217064],
       [ 492.38799154,  556.77439415,  409.99956099,  407.8930006 ,
         472.68349636,  631.49797654,  403.78552361,  633.68004476,
         567.5998023 ,  423.75540134],
       [ 350.1296884 ,  366.95278366,  496.76113889,  303.47862938,
         610.48187848,  276.91017664,  358.4482971 ,  530.03442731,
         331.05649739,  428.44382145],
       [ 302.98888858,  333.32300577,  348.19877722,  332.07587914,
         385.85714973,  581.79531664,  619.64409008,  420.18390534,
         543.04405937,  207.28757858],
       [ 523.43849515,  525.24403129,  400.84224262,  391.87383153,
         455.72001982,  403.02339669,  533.58746556,  480.31046377,
         566.23280914,  531.6247242 ],
       [ 546.72124222,  266.81594383,  504.1512776 ,  399.49133501,
         394.6952917 ,  582.93328833,  353.94079536,  383.81121652,
         299.87475231,  544.07238942],
       [ 430.91197558,  576.5203868 ,  286.03048542,  438.30897557,
         412.04042185,  353.10917543,  413.63772975,  647.14038542,
         591.25455248,  556.45493415],
       [ 334.92019753,  767.80389948,  381.982054  ,  234.44691863,
         391.31525815,  415.83189699,  336.99750123,  358.8042095 ,
         381.00677442,  488.71361319],
       [ 704.62956747,  563.13759423,  468.97808849,  366.55222553,
         352.5096905 ,  481.97065088,  391.47128438,  513.20375963,
         462.6762823 ,  242.05034098],
       [ 202.21029285,  367.3616573 ,  339.35849879,  448.42933906,
         545.00966775,  427.02261553,  582.71591752,  332.08112035,
         537.13389779,  448.60400089],
       [ 352.3626168 ,  370.17395382,  537.06004889,  360.70509766,
         463.74176473,  380.54643498,  402.61484566,  636.45985458,
         564.3316482 ,  440.59228808],
       [ 387.32355487,  357.56699865,  470.26784266,  487.17456386,
         490.82817128,  441.86735025,  380.3326202 ,  354.96961139,
         270.70095169,  292.08762615],
       [ 625.32154002,  464.92079704,  571.73110861,  480.93378021,
         399.57290238,  509.90861872,  373.5489584 ,  598.89819313,
         433.59308674,  488.8606805 ],
       [ 392.145978  ,  354.88786541,  540.75265742,  381.81679892,
         425.6351992 ,  362.17580821,  585.83057712,  511.53531085,
         653.8656794 ,  422.62056854],
       [ 232.81061492,  199.09341681,  529.57896022,  448.93288877,
         650.60507882,  530.00122388,  418.58174858,  529.24922256,
         470.39556704,  558.03710062],
       [ 495.79375113,  572.88253659,  413.35545587,  587.07280359,
         316.14414103,  443.69709127,  527.5582951 ,  517.65936524,
         476.80660548,  444.71044519],
       [ 271.75477577,  536.86008145,  408.810841  ,  491.06158182,
         392.34342951,  333.40552142,  435.98402952,  494.12765306,
         302.25867882,  413.16910196],
       [ 490.09581074,  654.15641547,  682.55637762,  516.74030117,
         533.25353608,  338.46981277,  368.4183214 ,  506.66488095,
         360.45027068,  377.36214527],
       [ 375.81469355,  476.89089893,  606.8629692 ,  518.68947468,
         467.74276059,  529.43810778,  498.40564011,  387.85758575,
         633.6556068 ,  445.90576367],
       [ 478.75637022,  502.83329922,  450.04894959,  216.72545155,
         599.93872877,  547.87948275,  415.38509935,  442.78702762,
         596.67699753,  524.21126813],
       [ 476.9277205 ,  356.55711545,  543.3025581 ,  423.16062654,
         474.64864836,  469.54505304,  456.13139638,  654.4820455 ,
         446.91303815,  432.77582977],
       [ 538.92672353,  442.13955141,  274.6244844 ,  443.04907538,
         298.87533339,  553.86182502,  534.10189334,  515.00164259,
         539.4139928 ,  617.8140593 ],
       [ 413.7268355 ,  398.93574269,  490.48668015,  428.99554965,
         419.53672363,  578.58683823,  447.49931945,  506.75696192,
         680.9650155 ,  390.91275773],
       [ 635.38883004,  606.62387273,  436.02445671,  456.31083202,
         562.3119456 ,  700.01303405,  528.83258492,  347.8618614 ,
         329.94310665,  524.72270227],
       [ 492.09063878,  437.79015883,  334.83244662,  423.81213355,
         514.58462846,  354.98881028,  496.54269628,  380.12079065,
         692.68634957,  131.92190412],
       [ 541.55462662,  486.58439816,  350.03725702,  343.92063402,
         518.45160613,  525.25232862,  500.30257393,  505.9165192 ,
         506.27762214,  558.51972707],
       [ 148.76076333,  359.30892236,  352.79721959,  359.75849769,
         565.75769694,  334.55845213,  308.24224061,  585.54618549,
         376.29403958,  297.30109537],
       [ 548.50524478,  408.22541498,  555.07723848,  574.53824632,
         364.96537099,  413.13696862,  272.37723348,  449.16127374,
         297.53516467,  597.15942575],
       [ 385.72806427,  534.71300808,  387.40318991,  651.15217971,
         291.23741164,  384.2158285 ,  315.09068148,  442.45961467,
         492.27647909,  460.69605376],
       [ 629.90550506,  343.950831  ,  301.3552114 ,  529.69630767,
         487.02992895,  331.85191798,  490.9980054 ,  472.38865483,
         400.74688824,  352.32038419],
       [ 517.93709105,  553.03631218,  400.54896818,  440.90917696,
         442.99725421,  793.65481379,  358.42170346,  360.83078563,
         538.4462266 ,  368.49753524],
       [ 433.24211466,  415.34884469,  459.94103542,  520.0266389 ,
         398.24921251,  341.16645897,  500.68633514,  348.70809112,
         416.59370821,  584.08917009],
       [ 560.89716988,  206.92628065,  471.89949533,  347.5433062 ,
         350.22630392,  365.63062474,  415.57318944,  523.97103689,
         438.24598122,  370.92473504],
       [ 328.53756503,  543.60268089,  409.228765  ,  491.48612273,
         345.5301853 ,  501.44253259,  509.53835998,  405.86788979,
         440.38751232,  403.25486474],
       [ 512.98340321,  542.13411813,  384.32797904,  387.20702584,
         526.65405353,  473.49530246,  424.72977929,  402.76052588,
         418.04609016,  293.54247962],
       [ 493.42547637,  445.14404001,  274.47062719,  596.01746299,
         594.45480972,  588.94045152,  432.9293811 ,  378.67675649,
         476.89453964,  692.10714854],
       [ 366.63429888,  403.84551935,  452.6876537 ,  608.19871786,
         630.29967233,  477.37191349,  460.2925561 ,  352.84800993,
         587.66833373,  533.66931725],
       [ 466.25915526,  362.95962698,  499.85218527,  494.03330708,
         517.5487912 ,  423.07948021,  501.7894371 ,  592.04494501,
         442.38605944,  361.78422153],
       [ 599.65545836,  502.59632122,  347.85646682,  536.34806351,
         524.99002225,  564.32573006,  597.54073454,  499.72179667,
         384.0447522 ,  357.66740199],
       [ 358.18364835,  614.55856299,  331.39749545,  544.96689132,
         387.95519893,  615.91556795,  407.39395531,  489.39168653,
         403.26857457,  443.29172626],
       [ 184.40429076,  337.95032486,  398.97442347,  651.46486733,
         474.16394589,  442.79188043,  471.39102921,  425.80834737,
         461.47353321,  405.39602659],
       [ 454.13908741,  454.987639  ,  600.81463963,  320.83575023,
         472.79988826,  328.59425693,  454.3570192 ,  471.06717183,
         468.69206013,  452.99588274],
       [ 397.96344693,  436.74684101,  405.53298363,  570.284488  ,
         574.11712298,  335.93003717,  617.50128227,  513.13196484,
         470.28696207,  409.31084751],
       [ 477.66303456,  175.25542754,  390.42799839,  475.04379466,
         425.05436348,  554.13137457,  585.63345916,  467.09842678,
         643.43489703,  467.54440723],
       [ 379.25207371,  440.59630172,  432.52563799,  531.3449845 ,
         433.56282293,  401.89999467,  523.46903089,  540.39581431,
         454.97304261,  593.69682356],
       [ 525.62434681,  500.4781448 ,  494.85280308,  253.04671084,
         343.2435028 ,  644.99356706,  431.55721944,  331.98020261,
         697.38278026,  450.62469306],
       [ 424.16997074,  569.15397041,  430.27006071,  419.31827785,
         526.10144916,  584.22185695,  448.46508265,  650.852166  ,
         642.39370537,  469.55896411],
       [ 465.29837453,  540.95881371,  444.14809838,  445.0962054 ,
         362.37844302,  258.74353808,  553.51502862,  490.18890815,
         441.69462936,  539.81971905],
       [ 604.58124468,  485.01384715,  436.1798742 ,  426.84935225,
         482.24847514,  497.68124061,  461.66380355,  521.21634603,
         471.30187429,  603.96281144]])

10）np.random.random(size=None)

In [74]:

np.random.random(size=10) # 生成0-1之间的浮点数

Out[74]:

array([ 0.67428546,  0.47222544,  0.20237801,  0.39959138,  0.13957155,
        0.0013156 ,  0.97975301,  0.38834253,  0.47289323,  0.58902421])

用随机数生成图片

In [78]:

boy = np.random.random((667,480,3))

plt.imshow(boy,cmap="gray")

plt.show()

二、ndarray的属性

数组的常用属性：

维度 ndim， 大小 size， 形状 shape， 元素类型 dtype， 每项大小 itemsize， 数据 data

In [79]:

tigger = plt.imread("./source/tigger.jpg")

tigger

Out[79]:

array([[[232, 228, 227],
        [232, 228, 227],
        [232, 228, 227],
        ..., 
        [221, 216, 212],
        [221, 216, 212],
        [221, 216, 212]],

       [[230, 226, 225],
        [231, 227, 226],
        [231, 227, 226],
        ..., 
        [221, 216, 212],
        [221, 216, 212],
        [221, 216, 212]],

       [[228, 224, 223],
        [229, 225, 224],
        [229, 225, 224],
        ..., 
        [221, 216, 212],
        [221, 216, 212],
        [221, 216, 212]],

       ..., 
       [[200, 195, 191],
        [207, 202, 198],
        [215, 210, 206],
        ..., 
        [159, 166, 182],
        [168, 175, 191],
        [157, 164, 180]],

       [[205, 200, 196],
        [202, 197, 193],
        [204, 199, 195],
        ..., 
        [147, 165, 179],
        [142, 161, 176],
        [136, 155, 170]],

       [[205, 200, 196],
        [202, 197, 193],
        [204, 199, 195],
        ..., 
        [147, 165, 179],
        [142, 161, 176],
        [136, 155, 170]]], dtype=uint8)

In [80]:

tigger.ndim

Out[80]:

3

In [81]:

girl.ndim

Out[81]:

3

In [84]:

nd

Out[84]:

array([1, 2, 3])

In [83]:

nd.size

Out[83]:

3

In [85]:

tigger.shape

Out[85]:

(786, 1200, 3)

In [88]:

nd.dtype

Out[88]:

dtype('int32')

In [89]:

tigger.dtype

Out[89]:

dtype('uint8')

In [86]:

nd.itemsize

Out[86]:

4

In [92]:

tigger.itemsize

Out[92]:

1

In [94]:

nd.data

Out[94]:

<memory at 0x000001C74AA22408>

In [93]:

tigger.data

Out[93]:

<memory at 0x000001C749E32A98>

三、ndarray的基本操作

1、索引

In [96]:

l = [1,3,4,6,7,8,0]

l[-1]

Out[96]:

0

In [98]:

nd = np.random.randint(0,4,size=4)

nd

Out[98]:

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

In [101]:

nd[0]

Out[101]:

1

In [104]:

lp = [[1,2,4],

    [4,5,6]

    ]

lp

Out[104]:

[[1, 2, 4], [4, 5, 6]]

In [107]:

lp[0][1]

Out[107]:

2

In [108]:

lp[0,1]

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-108-56bfe0f81cf9> in <module>()
----> 1 lp[0,1]

TypeError: list indices must be integers or slices, not tuple

In [ ]:

​

In [109]:

nd = np.random.randint(0,10,size=(6,6))

nd

Out[109]:

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

In [110]:

nd[0]

Out[110]:

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

In [111]:

nd[0][1]

Out[111]:

7

In [112]:

nd[0,1] # 先访问第0个维度的第0个，再访问第1个维度中第1

Out[112]:

7

In [113]:

nd[1,2,3] # 维度不足3个，不这样访问

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-113-e998ed9d4f35> in <module>()
----> 1 nd[1,2,3]

IndexError: too many indices for array

In [114]:

nd[1,4] = 20000

nd

Out[114]:

array([[    2,     7,     2,     9,     3,     3],
       [    6,     8,     7,     6, 20000,     0],
       [    5,     8,     2,     5,     8,     5],
       [    2,     7,     1,     2,     9,     6],
       [    3,     7,     1,     0,     1,     2],
       [    1,     9,     0,     0,     7,     8]])

In [116]:

nd

Out[116]:

array([[    2,     7,     2,     9,     3,     3],
       [    6,     8,     7,     6, 20000,     0],
       [    5,     8,     2,     5,     8,     5],
       [    2,     7,     1,     2,     9,     6],
       [    3,     7,     1,     0,     1,     2],
       [    1,     9,     0,     0,     7,     8]])

In [118]:

nd[[1,1,1,1,1]]

nd[[1,4,5,2,0]] # 连续访问多个，可以把你要访问的那些下标写成一个列表

Out[118]:

array([[    6,     8,     7,     6, 20000,     0],
       [    3,     7,     1,     0,     1,     2],
       [    1,     9,     0,     0,     7,     8],
       [    5,     8,     2,     5,     8,     5],
       [    2,     7,     2,     9,     3,     3]])

2、切片

In [120]:

nd = np.random.randint(0,10,size=(4,4))

nd

Out[120]:

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

In [126]:

nd[0:5] # 区间左闭右开

Out[126]:

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

In [129]:

l[0:10]

Out[129]:

[1, 3, 4, 6, 7, 8, 0]

In [130]:

nd[1:]

Out[130]:

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

In [134]:

nd[::-1]

Out[134]:

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

In [135]:

nd

Out[135]:

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

In [138]:

nd[:,0:2]

Out[138]:

array([[7, 5],
       [7, 5],
       [5, 9],
       [2, 1]])

把girl调头

In [139]:

girl

Out[139]:

array([[[225, 231, 231],
        [229, 235, 235],
        [222, 228, 228],
        ..., 
        [206, 213, 162],
        [211, 213, 166],
        [217, 220, 173]],

       [[224, 230, 230],
        [229, 235, 235],
        [223, 229, 229],
        ..., 
        [206, 213, 162],
        [211, 213, 166],
        [217, 220, 173]],

       [[224, 230, 230],
        [229, 235, 235],
        [223, 229, 229],
        ..., 
        [206, 213, 162],
        [211, 213, 166],
        [219, 221, 174]],

       ..., 
       [[175, 187, 213],
        [180, 192, 218],
        [175, 187, 213],
        ..., 
        [155, 162, 180],
        [153, 160, 178],
        [156, 163, 181]],

       [[175, 187, 213],
        [180, 192, 218],
        [174, 186, 212],
        ..., 
        [155, 162, 180],
        [153, 160, 178],
        [155, 162, 180]],

       [[177, 189, 215],
        [181, 193, 219],
        [174, 186, 212],
        ..., 
        [155, 162, 180],
        [153, 160, 178],
        [156, 163, 181]]], dtype=uint8)

In [145]:

girl2 = girl[::-3]

In [146]:

plt.imshow(girl2)

plt.show()

In [166]:

girl3 = girl[::2,::-2]

plt.imshow(girl3)

plt.show()

In [167]:

girl4 = girl[:,:,::1]

girl4[:,:,:2] = 0

plt.imshow(girl4)

plt.show()

In [ ]:

​

3、变形

reshape()

resize()

In [170]:

tigger

Out[170]:

array([[[232, 228, 227],
        [232, 228, 227],
        [232, 228, 227],
        ..., 
        [221, 216, 212],
        [221, 216, 212],
        [221, 216, 212]],

       [[230, 226, 225],
        [231, 227, 226],
        [231, 227, 226],
        ..., 
        [221, 216, 212],
        [221, 216, 212],
        [221, 216, 212]],

       [[228, 224, 223],
        [229, 225, 224],
        [229, 225, 224],
        ..., 
        [221, 216, 212],
        [221, 216, 212],
        [221, 216, 212]],

       ..., 
       [[200, 195, 191],
        [207, 202, 198],
        [215, 210, 206],
        ..., 
        [159, 166, 182],
        [168, 175, 191],
        [157, 164, 180]],

       [[205, 200, 196],
        [202, 197, 193],
        [204, 199, 195],
        ..., 
        [147, 165, 179],
        [142, 161, 176],
        [136, 155, 170]],

       [[205, 200, 196],
        [202, 197, 193],
        [204, 199, 195],
        ..., 
        [147, 165, 179],
        [142, 161, 176],
        [136, 155, 170]]], dtype=uint8)

In [171]:

tigger.shape

Out[171]:

(786, 1200, 3)

In [172]:

nd = np.random.randint(0,20,size=12)

nd (1*12)

Out[172]:

array([ 2,  1,  0, 17, 18, 13, 11,  7,  8, 19,  2, 16])

In [173]:

nd.reshape((3,4))

Out[173]:

array([[ 2,  1,  0, 17],
       [18, 13, 11,  7],
       [ 8, 19,  2, 16]])

In [174]:

nd

Out[174]:

array([ 2,  1,  0, 17, 18, 13, 11,  7,  8, 19,  2, 16])

In [175]:

nd.reshape((4,4)) # 元素总数不一致，是不能相互转变的

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-175-133d72e119fa> in <module>()
----> 1 nd.reshape((4,4))

ValueError: cannot reshape array of size 12 into shape (4,4)

In [176]:

nd.reshape((1,6))

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-176-706a52b78a91> in <module>()
----> 1 nd.reshape((1,6))

ValueError: cannot reshape array of size 12 into shape (1,6)

In [178]:

nd.reshape((1,1,2,6))

Out[178]:

array([[[[ 2,  1,  0, 17, 18, 13],
         [11,  7,  8, 19,  2, 16]]]])

In [179]:

nd.resize((3,4))

In [180]:

nd

Out[180]:

array([[ 2,  1,  0, 17],
       [18, 13, 11,  7],
       [ 8, 19,  2, 16]])

【注意】

1、resize()和reshape()都是对数组进行变形，resize()是在原来数组上直接变形，reshape()生成一个新的数组

2、变形的时候，新数组的总元素个数要和原来的保持一致

拼图游戏

In [181]:

girl = plt.imread("./source/girl2.jpg")

plt.imshow(girl)

plt.show()

In [186]:

tigger[150:450,200:500] = girl

plt.imshow(tigger)

plt.show()

In [ ]:

​

4、级联

In [4]:

nd1 = np.random.randint(0,100,size=(4,4))

nd2 = np.random.randint(0,100,size=(4,3))

print(nd1)

print(nd2)

[[86 87 27  4]
 [30 93 58 74]
 [37 83 74 48]
 [ 9  4 81 19]]
[[10 36 77]
 [41  5 32]
 [38 30  6]
 [17 20 39]]

级联：就是按照指定的维度把两个数组链接在一起

级联注意点：

1、参与级联的数组要构成一个列表(或元组)

2、维度必须一样

3、形状必须相符（如：二维数组axis=0时，列数必须一样，axis=1的时候行数必须一样）

4、级联的默认方向shape的第一个值所代表的那个维度（axis默认为0）

In [8]:

np.concatenate([nd1,nd2],axis=1) 

# 参数一，参与级联的那些数组；参数二，级联的维度

Out[8]:

array([[86, 87, 27,  4, 10, 36, 77],
       [30, 93, 58, 74, 41,  5, 32],
       [37, 83, 74, 48, 38, 30,  6],
       [ 9,  4, 81, 19, 17, 20, 39]])

In [13]:

nd3 = np.random.randint(0,100,size=(6,4))

nd3

Out[13]:

array([[55, 91,  5, 82],
       [95,  0, 17, 78],
       [91, 89, 27, 35],
       [37, 82, 78, 61],
       [ 6, 58, 14, 62],
       [21, 62, 45, 35]])

In [14]:

np.concatenate([nd1,nd3])

Out[14]:

array([[86, 87, 27,  4],
       [30, 93, 58, 74],
       [37, 83, 74, 48],
       [ 9,  4, 81, 19],
       [55, 91,  5, 82],
       [95,  0, 17, 78],
       [91, 89, 27, 35],
       [37, 82, 78, 61],
       [ 6, 58, 14, 62],
       [21, 62, 45, 35]])

hstack()和vstack()

hstack()把列数组改成一个行数组（把所有行上数组拼接成一个一行的数组）

vstack()把一个单行的数组拆成多行（每个行上只有一个数字）

In [16]:

nd3 = np.random.randint(0,10,size=(10,1))

nd3

Out[16]:

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

In [18]:

np.hstack(nd3)

Out[18]:

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

In [19]:

nd4 = np.random.randint(0,50,size=(4,3))

nd4

Out[19]:

array([[16, 41, 36],
       [45, 18,  3],
       [37,  9,  0],
       [43, 18, 46]])

In [20]:

np.hstack(nd4)

Out[20]:

array([16, 41, 36, 45, 18,  3, 37,  9,  0, 43, 18, 46])

In [21]:

np.vstack(nd4)

Out[21]:

array([[16, 41, 36],
       [45, 18,  3],
       [37,  9,  0],
       [43, 18, 46]])

In [22]:

nd5 = np.random.randint(0,10,size=10)

nd5

Out[22]:

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

In [23]:

np.vstack(nd5)

Out[23]:

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

5、切分

split()

vsplit()

hsplit()

In [28]:

nd = np.random.randint(0,100,size=(5,6))

nd

Out[28]:

array([[63, 92, 23, 12, 35, 69],
       [65, 30, 73, 78, 36, 70],
       [56,  9, 16, 53, 89, 23],
       [57, 79, 35, 68, 68, 16],
       [19,  0, 30, 77, 81, 20]])

In [27]:

np.hsplit(nd,[1,3]) # 代表在水平方向切分

# 参数一，被切的数组  参数二，切分点（列表，在切分点前切分）

Out[27]:

[array([[54],
        [44],
        [32],
        [37],
        [69]]), array([[51, 80],
        [ 1, 90],
        [65,  0],
        [ 0, 52],
        [64, 32]]), array([[37, 40, 25],
        [91, 88,  6],
        [ 8, 63, 78],
        [44, 48, 29],
        [19,  4, 17]])]

In [29]:

np.vsplit(nd,[1,4]) # 纵向上切分 参数和横向类似

Out[29]:

[array([[63, 92, 23, 12, 35, 69]]), array([[65, 30, 73, 78, 36, 70],
        [56,  9, 16, 53, 89, 23],
        [57, 79, 35, 68, 68, 16]]), array([[19,  0, 30, 77, 81, 20]])]

In [30]:

np.split(nd,[1,3],axis=0) # 切行

Out[30]:

[array([[63, 92, 23, 12, 35, 69]]), array([[65, 30, 73, 78, 36, 70],
        [56,  9, 16, 53, 89, 23]]), array([[57, 79, 35, 68, 68, 16],
        [19,  0, 30, 77, 81, 20]])]

In [32]:

np.split(nd,(1,3),axis=1) # 切列

Out[32]:

[array([[63],
        [65],
        [56],
        [57],
        [19]]), array([[92, 23],
        [30, 73],
        [ 9, 16],
        [79, 35],
        [ 0, 30]]), array([[12, 35, 69],
        [78, 36, 70],
        [53, 89, 23],
        [68, 68, 16],
        [77, 81, 20]])]

6、副本

In [33]:

nd = np.random.randint(0,100,size=6)

nd

Out[33]:

array([13, 89, 33, 62, 74, 55])

In [34]:

nd1 = nd 

In [36]:

nd1[0] = 1000

nd1

Out[36]:

array([1000,   89,   33,   62,   74,   55])

In [38]:

nd # 此时nd1和nd变量引用是同一个数组对象（nd和nd1中存储的地址都是一样的）

# 在某种意义上这也是一种浅拷贝

Out[38]:

array([1000,   89,   33,   62,   74,   55])

用copy()函数对数组对象创建副本

In [40]:

nd2 = nd.copy()

nd2

Out[40]:

array([1000,   89,   33,   62,   74,   55])

In [41]:

nd2[2] = 2000

nd2

Out[41]:

array([1000,   89, 2000,   62,   74,   55])

In [42]:

nd

Out[42]:

array([1000,   89,   33,   62,   74,   55])

In [43]:

nd1

Out[43]:

array([1000,   89,   33,   62,   74,   55])

用列表创建数组的过程有木有创建副本

In [44]:

l = [1,2,3,4]

nd = np.array(l) # 把l拷贝了一份放在了数组nd中

In [45]:

nd[0] = 100

nd

Out[45]:

array([100,   2,   3,   4])

In [46]:

l

Out[46]:

[1, 2, 3, 4]

In [ ]:

​

四、ndarray的聚合操作

1、求和

In [55]:

nd = np.random.randint(0,10,size=(3,4))

nd

Out[55]:

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

In [56]:

np.sum(nd,axis=0)  # 把行按照列标对应加起来

Out[56]:

array([14, 10,  7, 21])

In [57]:

np.sum(nd,axis=1) # 把列按照对应行标加起来

Out[57]:

array([19, 20, 13])

In [59]:

np.sum(nd) # 把所有的元素加起来

Out[59]:

52

In [60]:

nd.sum()

Out[60]:

52

2、最值

In [69]:

nd

Out[69]:

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

In [63]:

np.max(nd)

Out[63]:

9

In [65]:

np.max(nd,axis=0) # 求每个列里面最大值

Out[65]:

array([8, 8, 4, 9])

In [66]:

np.max(nd,axis=1) # 求每一行里面的最大值

Out[66]:

array([9, 8, 6])

In [67]:

np.argmax(nd)

Out[67]:

3

In [68]:

np.argmin(nd)

Out[68]:

2

In [71]:

np.argmax(nd,axis=0) # 在列上求最大值并返回最大值的行标

Out[71]:

array([1, 0, 1, 0], dtype=int64)

In [72]:

np.argmin(nd,axis=1) # 在每个行上求最小值返回列标

Out[72]:

array([2, 1, 1], dtype=int64)

3、其他聚合操作

Function Name    NaN-safe Version    Description
np.sum    np.nansum    Compute sum of elements
np.prod    np.nanprod    Compute product of elements
np.mean    np.nanmean    Compute mean of elements
np.std    np.nanstd    Compute standard deviation
np.var    np.nanvar    Compute variance
np.min    np.nanmin    Find minimum value
np.max    np.nanmax    Find maximum value
np.argmin    np.nanargmin    Find index of minimum value
np.argmax    np.nanargmax    Find index of maximum value
np.median    np.nanmedian    Compute median of elements
np.percentile    np.nanpercentile    Compute rank-based statistics of elements
np.any    N/A    Evaluate whether any elements are true
np.all    N/A    Evaluate whether all elements are true
np.power 幂运算

In [73]:

nd = np.array([1,2,3,np.nan])

nd

Out[73]:

array([  1.,   2.,   3.,  nan])

In [76]:

np.nan + 100

np.nan*100

# nan表示缺失（默认是浮点型），缺失和任何数做运算结果都是缺失

Out[76]:

nan

In [74]:

nd.sum()

Out[74]:

nan

In [77]:

np.nansum(nd) # 以nan开头的聚合方法，可在聚合运算的直接把nan剔除

Out[77]:

6.0

思考题：给定一个4维矩阵，如何得到最后两维的和？

In [78]:

nd = np.random.randint(0,10,size=(2,3,2,2))

nd

Out[78]:

array([[[[4, 9],
         [5, 1]],

        [[2, 1],
         [7, 5]],

        [[2, 7],
         [4, 9]]],


       [[[8, 9],
         [6, 4]],

        [[7, 3],
         [5, 4]],

        [[5, 4],
         [7, 6]]]])

In [81]:

nd.sum(axis=-1)

nd.sum(axis=-2)

nd.sum(axis=(-1,-2))

Out[81]:

array([[19, 15, 22],
       [27, 19, 22]])

思考题：如何根据第3列来对一个5*5矩阵排序？

In [84]:

nd = np.random.randint(0,20,size=(5,5))

nd

Out[84]:

array([[ 5, 19,  1,  4, 12],
       [13, 16, 17, 18, 17],
       [19, 12,  4, 19, 10],
       [16, 11, 18,  3, 14],
       [14,  0, 15, 15, 10]])

In [85]:

nd[[3,0,4,1,2]] #??

Out[85]:

array([[16, 11, 18,  3, 14],
       [ 5, 19,  1,  4, 12],
       [14,  0, 15, 15, 10],
       [13, 16, 17, 18, 17],
       [19, 12,  4, 19, 10]])

In [92]:

nd

Out[92]:

array([[ 1,  4,  5, 12, 19],
       [13, 16, 17, 17, 18],
       [ 4, 10, 12, 19, 19],
       [ 3, 11, 14, 16, 18],
       [ 0, 10, 14, 15, 15]])

In [94]:

sort_ind = np.argsort(nd[:,3]) # 排序并且返回排序以后的下标序列

sort_ind

Out[94]:

array([0, 4, 3, 1, 2], dtype=int64)

In [95]:

nd[sort_ind]

. . .

思路

In [ ]:

​

五、ndarray的矩阵操作

1. 基本矩阵操作

1）算术运算（即加减乘除）

In [96]:

nd = np.random.randint(0,10,size=(5,5))

nd

Out[96]:

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

In [97]:

nd + 3  # 加运算，启动广播机制把缺失的地方补全

Out[97]:

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

In [98]:

nd*2 # 乘法不需要广播机制

Out[98]:

array([[18,  0,  8,  4, 14],
       [14,  6, 18, 12, 12],
       [10, 18, 12,  6,  0],
       [ 8, 10, 10, 16,  6],
       [ 4, 10,  6,  6,  6]])

In [99]:

nd/2

Out[99]:

array([[ 4.5,  0. ,  2. ,  1. ,  3.5],
       [ 3.5,  1.5,  4.5,  3. ,  3. ],
       [ 2.5,  4.5,  3. ,  1.5,  0. ],
       [ 2. ,  2.5,  2.5,  4. ,  1.5],
       [ 1. ,  2.5,  1.5,  1.5,  1.5]])

2）矩阵积

In [103]:

nd1 = np.random.randint(0,5,size=(4,3))

nd2 = np.random.randint(0,5,size=(3,4))

print(nd1)

print(nd2)

. . .

In [104]:

np.dot(nd1,nd2)

. . .

2. 广播机制

ndarray的广播机制的两条规则：

• 1、为缺失维度补1
• 2、假定缺失的元素用已有值填充

In [105]:

m = np.random.randint(0,10,size=(2,3))

m

Out[105]:

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

In [106]:

a = np.arange(3)

a

Out[106]:

array([0, 1, 2])

In [107]:

# 求m+a

m+a

Out[107]:

array([[ 3,  7, 11],
       [ 4,  8,  9]])

In [108]:

a = np.arange(3).reshape((3,1))

a

Out[108]:

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

In [109]:

b = np.arange(3)

b

Out[109]:

array([0, 1, 2])

In [110]:

a+b

Out[110]:

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

0 0 0 0 1 2

1 1 1 + 0 1 2

2 2 2 0 1 2

In [111]:

a + 3

Out[111]:

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

In [112]:

a = np.random.randint(0,5,size=(3,3))

a

Out[112]:

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

In [113]:

b = np.random.randint(0,5,size=(2,2))

b

Out[113]:

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

In [114]:

a + b # 如果行和列个数均不相同无法进行补全，不能相加

-------------------------------------------------------------
ValueError                  Traceback (most recent call last)
<ipython-input-114-f96fb8f649b6> in <module>()
----> 1 a + b

ValueError: operands could not be broadcast together with shapes (3,3) (2,2) 

排序

In [120]:

nd = np.random.randint(0,100,size=(10))

nd

Out[120]:

array([70, 90, 72, 23, 55, 85, 24, 88, 91, 13])

In [119]:

nd.sort() # 改变原来的数组

In [118]:

nd

Out[118]:

array([13, 26, 35, 43, 46, 59, 76, 77, 81, 94])

In [121]:

np.sort(nd)# 不会改变原来的数组，而是生成一个新的数组

Out[121]:

array([13, 23, 24, 55, 70, 72, 85, 88, 90, 91])

In [122]:

nd

Out[122]:

array([70, 90, 72, 23, 55, 85, 24, 88, 91, 13])

In [123]:

nd1 = np.random.randint(0,100,size=(4,4))

nd1

Out[123]:

array([[95, 82, 52, 15],
       [70, 87,  2, 54],
       [52, 66, 26, 95],
       [ 9, 49, 63, 48]])

In [126]:

np.sort(nd1,axis=1) # axis=1排的是列标 axis=0排的是行标

Out[126]:

array([[15, 52, 82, 95],
       [ 2, 54, 70, 87],
       [26, 52, 66, 95],
       [ 9, 48, 49, 63]])

In [127]:

np.sort(nd1,axis=0)

Out[127]:

array([[ 9, 49,  2, 15],
       [52, 66, 26, 48],
       [70, 82, 52, 54],
       [95, 87, 63, 95]])

In [128]:

# argsort()排的是下标

np.argsort(nd1,axis=0)

Out[128]:

array([[3, 3, 1, 0],
       [2, 2, 2, 3],
       [1, 0, 0, 1],
       [0, 1, 3, 2]], dtype=int64)

In [ ]:

​