NumPy基础：多维数组对象

• 创建ndarray
  • array：将输入数据转换为ndarry
  • arange：类似于range，返回一个ndarray
  • ones、ones_like：根据指定形状和dtype创建一个全1数组。ones_like以另一个数组为参照，根据其形状和dtype创建一个全1数组
  • zeros、zeros_like：根据指定形状和dtype创建一个全0数组。zeros_like以另一个数组为参照，根据其形状和dtype创建一个全0数组
  • empty、empty_like：创建新数组，只分配内存空间但不填充任何值
  • eye、identity：创建一个正方的N*N单位矩阵（对角线为1，其余为0）

import numpy as np
data1 = np.array([1,9,8,8,0,4,1,6])
print(data1)
'''
[1 9 8 8 0 4 1 6]
'''

data2 = np.array([[1,2,3],[4,5,6]])
print(data2)
'''
[[1 2 3]
 [4 5 6]]
 '''
data3 = np.arange(10)
print(data3)
'''
[0 1 2 3 4 5 6 7 8 9]
'''
data4 = np.ones(3)
print(data4)
'''
[1. 1. 1.]
'''
data_like = [[1,2,3],[4,5,6]]
data5 = np.ones_like(data_like)
print(data5)
'''
[[1 1 1]
 [1 1 1]]
'''
data4 = np.ones(3)
print(data4)
'''
[1. 1. 1.]
'''
data6 = np.zeros(3)
print(data6)
'''
[0. 0. 0.]
'''
data_like = [[1,2,3],[4,5,6]]
data7 = np.zeros_like(data_like)
print(data7)
'''
[[0 0 0]
 [0 0 0]]
'''
data8 = np.empty(3)
print(data8)
'''
[4.24399158e-314 8.48798317e-314 1.27319747e-313]
'''
data9 = np.eye(3)
print(data9)
'''
[[1. 0. 0.]
 [0. 1. 0.]
 [0. 0. 1.]]
'''

• ndarray的数据类型

import numpy as np

# np.array 会尝试为新建的数组推断一个较为合适的数据类型
arr1 = np.array([1,2,3])
print(arr1.dtype)
'''
int32
'''

# 可使用dtype数据类型
arr2 = np.array([1,2,3],dtype=np.float)
print(arr2.dtype)
'''
float64
'''

arr3 = np.array([1.1,2,3])
print(arr3.dtype)
'''
float64
'''

arr4 = np.array([1.1,2,3],dtype=np.int32)
print(arr4.dtype)
'''
int32
'''

# 通过astype方法可转换dtype
# 调用astype会创建出一个新数组（原始数据的一份拷贝）
arr5 = np.array([1,2,3])
float_arr = arr5.astype(np.float64)
print(float_arr.dtype)
'''
float64
'''

# 注意：如果浮点数转换为整数，则小数部分将会被截断
arr6 = np.array([1.2,2,3])
int_arr = arr6.astype(np.int32)
print(int_arr.dtype)
'''
int32
'''
print(int_arr)
'''
[1 2 3]
'''

• 数组和标量之间的运算

import numpy as np

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

print(arr+arr)
'''
[[ 2  4  6]
 [ 8 10 12]]
'''
print(arr-arr)
'''
[[0 0 0]
 [0 0 0]]
'''
print(arr*arr)
'''
[[ 1  4  9]
 [16 25 36]]
'''
print(arr/arr)
'''
[[1. 1. 1.]
 [1. 1. 1.]]
'''
print(arr**arr)
'''
[[    1     4    27]
 [  256  3125 46656]]
'''
print(arr+1)
'''
[[2 3 4]
 [5 6 7]]
'''

• 基本的索引和切片

import numpy as np

# 一维数组
arr = np.arange(10)
print(arr)
'''
[0 1 2 3 4 5 6 7 8 9]
'''
print(arr[5])
'''
5
'''
print(arr[5:8])
'''
[5 6 7]
'''
# 当你将一个标量赋值给一个切片时，该值会自动传播到整个选区
arr[5:8] = 12
print(arr)
'''
[ 0  1  2  3  4 12 12 12  8  9]
'''
# 数组切片是原始数据视图。这意味着数据不会被复制，任何修改直接反应到源数组上
arr_slice = arr[5:8]
arr_slice[1] = 12345
print(arr)
'''
[    0     1     2     3     4    12 12345    12     8     9]
'''
arr_slice[:] = 64
print(arr)
'''
[ 0  1  2  3  4 64 64 64  8  9]
'''

# 二维数组
arr2d = np.array([[1,2,3],[4,5,6],[7,8,9]])
print(arr2d)
'''
[[1 2 3]
 [4 5 6]
 [7 8 9]]
'''
print(arr2d[2])
'''
[7 8 9]
'''
print(arr2d[0][2])
'''
3
'''
print(arr2d[0,2])
'''
3
'''
# 多维数组
arr3d = np.array([[[1,2,3],[4,5,6]],[[7,8,9],[10,11,12]]])
print(arr3d)
'''
[[[ 1  2  3]
  [ 4  5  6]]

 [[ 7  8  9]
  [10 11 12]]]
'''
print(arr3d[0])
'''
[[1 2 3]
 [4 5 6]]
'''
old_values = arr3d[0].copy()
arr3d[0] = 42
print(arr3d)
'''
[[[42 42 42]
  [42 42 42]]

 [[ 7  8  9]
  [10 11 12]]]
'''
arr3d[0] = old_values
print(arr3d)
'''
[[[ 1  2  3]
  [ 4  5  6]]

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

'''
print(arr3d[1,0])
'''
[7 8 9]
'''

• 切片索引

import numpy as np

# 一维数组
arr = np.arange(10)
print(arr[1:6])
'''
[1 2 3 4 5]
'''

# 二维数组
arr2d = np.array([[1,2,3],[4,5,6],[7,8,9]])
print(arr2d[:2])
'''
[[1 2 3]
 [4 5 6]]
'''
print(arr2d[:2,1:])
'''
[[2 3]
 [5 6]]
'''
# 索引与切片混合
print(arr2d[1,:2])
'''
[4 5]
'''
# 只有冒号表示选区整个轴
print(arr2d[:,:2])
'''
[[1 2]
 [4 5]
 [7 8]]
'''
# 对切片赋值也会扩散到整个选区
arr2d[:,:2]=0
print(arr2d)
'''
[[0 0 3]
 [0 0 6]
 [0 0 9]]
'''

• 布尔型索引

import numpy as np
from numpy.matlib import randn

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

# 等于
print(names=="Bob")
'''
[ True False False  True False False False]
'''
# 不等于
print(names!="Bob")
'''
[False  True  True False  True  True  True]
'''
# 组合多个条件使用&、|
print((names=="Bob")|(names=="Will"))
'''
[ True False  True  True  True False False]
'''
data = randn(7,3)
print(data)
'''
[[ 0.35234481  0.68539956  0.2206396 ]
 [-1.3719165  -0.42694698  1.28509104]
 [-0.95479498 -0.65378008 -0.1673056 ]
 [-1.79677508  0.18923784  1.67064335]
 [-1.24383276 -0.50056086 -0.7917794 ]
 [-0.92646918  0.47489349 -0.62463223]
 [ 0.0995606  -1.20420049 -1.55692415]]
'''
mask =(names=="Bob")  # 0和3为Ture，取0和3行
print(data[mask])
'''
[[ 0.35234481  0.68539956  0.2206396 ]
 [-1.79677508  0.18923784  1.67064335]]
'''
#
mask =(names=="Bob")|(names=="Will") # 0、2、3、4为Ture，取这4行
print(data[mask])
'''
[[ 0.35234481  0.68539956  0.2206396 ]
 [-0.95479498 -0.65378008 -0.1673056 ]
 [-1.79677508  0.18923784  1.67064335]
 [-1.24383276 -0.50056086 -0.7917794 ]]
'''
# 将data中负值置为0
data[data<0] = 0
print(data)
'''
[[0.35234481 0.68539956 0.2206396 ]
 [0.         0.         1.28509104]
 [0.         0.         0.        ]
 [0.         0.18923784 1.67064335]
 [0.         0.         0.        ]
 [0.         0.47489349 0.        ]
 [0.0995606  0.         0.        ]]
'''
# 通过布尔数组设置整条或整列的值
data[names=="Bob"] = 7
print(data)
'''
[[7.         7.         7.        ]
 [0.         0.5615304  0.        ]
 [0.         1.05144009 0.04887547]
 [7.         7.         7.        ]
 [0.5067665  0.         0.        ]
 [2.0465758  0.78871507 0.68937188]
 [1.58986486 0.86841601 1.46533603]]
'''

• 花式索引

import numpy as np

# 花式索引：利用整数数组进行索引
'''
arr = np.empty((8,4))
for i in range(8):
    arr[i] = i
'''
arr = np.arange(16).reshape((4,4))
print(arr)
'''
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]
 [12 13 14 15]]
'''
# 为了特定顺序选取行子集，需传入一个用于指定顺序的整数列
print(arr[[1,0,2]])
'''
[[ 4  5  6  7]
 [ 0  1  2  3]
 [ 8  9 10 11]]
'''
# 传入多个索引
print(arr[[1,0,2],[0,1,2]]) 
'''
[ 4  1 10]
'''
# 使用负数索引将会从末尾开始选行
print(arr[[-3,-2,-1]])
'''
[[ 4  5  6  7]
 [ 8  9 10 11]
 [12 13 14 15]]
'''

• 数组的转置与轴对换

import numpy as np

arr = np.arange(15).reshape((3,5))
print(arr)
'''
[[ 0  1  2  3  4]
 [ 5  6  7  8  9]
 [10 11 12 13 14]]
'''
print(arr.T)
'''
[[ 0  5 10]
 [ 1  6 11]
 [ 2  7 12]
 [ 3  8 13]
 [ 4  9 14]]
'''
# 矩阵内积
print(np.dot(arr.T,arr))
'''
[[125 140 155 170 185]
 [140 158 176 194 212]
 [155 176 197 218 239]
 [170 194 218 242 266]
 [185 212 239 266 293]]
'''

arr2 = np.arange(16).reshape((2,2,4))
print(arr2)
'''
[[[ 0  1  2  3]
  [ 4  5  6  7]]

 [[ 8  9 10 11]
  [12 13 14 15]]]
'''
print(arr2.transpose((1,0,2))) #行列索引值对换
'''
[[[ 0  1  2  3]
  [ 8  9 10 11]]

 [[ 4  5  6  7]
  [12 13 14 15]]]
'''

arr3 = np.arange(16).reshape((2,2,4))
print(arr3)
'''
[[[ 0  1  2  3]
  [ 4  5  6  7]]

 [[ 8  9 10 11]
  [12 13 14 15]]]
'''
print(arr3.swapaxes(1,2))
'''
[[[ 0  4]
  [ 1  5]
  [ 2  6]
  [ 3  7]]

 [[ 8 12]
  [ 9 13]
  [10 14]
  [11 15]]]
'''