python 2-5（2019-11-01 ）（上）

import numpy as np

x = np.arange(10)
x

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

X = np.arange(15).reshape(5,3)
X

array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11],
       [12, 13, 14]])

访问

x[2]

X[1,1]

切片

x[5:]

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

# 行切片,列切片
X[2:4,1:]

array([[ 7,  8],
       [10, 11]])

X[2:,0:2]

array([[ 6,  7],
       [ 9, 10],
       [12, 13]])

# 如果说不关心reshape的定外一个参数,我们可以写成-1,numpy指定推导出这个参数
x.reshape(5,-1)

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

numpy的与运算

array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11],
       [12, 13, 14]])

X+1

array([[ 1,  2,  3],
       [ 4,  5,  6],
       [ 7,  8,  9],
       [10, 11, 12],
       [13, 14, 15]])

X*2

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

np.sin(X)

array([[ 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]])

NUMPY 中的argsort()

x = np.arange(16)
x

array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15])

# 随机打乱
# from random import shuffle
np.random.shuffle(x)

array([ 6,  2,  7, 10, 13,  3, 11, 14,  1,  0,  5,  4,  9, 12, 15,  8])

np.argsort(x)

array([ 9,  8,  1,  5, 11, 10,  0,  2, 15, 12,  3,  6, 13,  4,  7, 14],
      dtype=int64)

Numpy中的布尔索引

names = np.array(['Bob','Joe','Will', 'Will','Bob','Joe','Will'])
names

array(['Bob', 'Joe', 'Will', 'Will', 'Bob', 'Joe', 'Will'], dtype='<U4')

# 使用np.random模块的randn生成一些正太分部的随即数据
data = np.random.randn(7,4)
data

array([[ 0.03290805,  0.90771856, -0.81345972,  0.46827413],
       [ 0.52700879,  0.33853328, -1.20246958,  1.14056211],
       [ 0.2077111 , -0.85287939,  0.64753495,  0.51345487],
       [ 0.56489484,  0.15367868,  0.38281581,  0.98537961],
       [ 0.29369419,  1.19652407,  0.18914402, -0.25025609],
       [-0.58819709, -1.03515838, -0.13233965, -0.06189865],
       [-1.90031352,  0.81946973, -0.68476672,  0.1286769 ]])

# 假设每个名字对应data数组的一行
names =='Bob'

array([ True, False, False, False,  True, False, False])

# 布尔型索引可以应用于数据的筛选
data[names == 'Bob']

array([[ 2.12140159,  0.265584  , -0.73249615,  0.31393966],
       [-0.05550636, -0.0832018 , -1.27019383, -0.18444114]])

# 布尔型索引应用于修改值
#选取所有JOE的行,并且全部值赋值为 0
data[names == 'Joe'] = 0
data

array([[ 2.12140159,  0.265584  , -0.73249615,  0.31393966],
       [ 0.        ,  0.        ,  0.        ,  0.        ],
       [-1.1163393 , -0.41695486, -0.48934225, -0.0688788 ],
       [ 0.41426097, -0.87294359,  0.33790829, -1.75023062],
       [-0.05550636, -0.0832018 , -1.27019383, -0.18444114],
       [ 0.        ,  0.        ,  0.        ,  0.        ],
       [ 1.37574555,  0.74651564, -1.61976615,  0.04435584]])

#选取所有Will的行,并将选取数据的后两列值赋值为 0
data[names == 'Will',2:] = 0
data

array([[ 0.03290805,  0.90771856, -0.81345972,  0.46827413],
       [ 0.52700879,  0.33853328, -1.20246958,  1.14056211],
       [ 0.2077111 , -0.85287939,  0.        ,  0.        ],
       [ 0.56489484,  0.15367868,  0.        ,  0.        ],
       [ 0.29369419,  1.19652407,  0.18914402, -0.25025609],
       [-0.58819709, -1.03515838, -0.13233965, -0.06189865],
       [-1.90031352,  0.81946973,  0.        ,  0.        ]])

最后编辑于：2019.11.01 23:40:20