numpy 数据分析基础
创建数组
| 方法 | 说明 |
|---|---|
| np.array( x ) | 将输入数据转化为一个ndarray| |
| np.array( x, dtype ) | 将输入数据转化为一个类型为type的ndarray| |
| np.asarray( array ) | 将输入数据转化为一个新的（copy） |
| np.ones( N ) | 生成一个N长度的一维全一ndarray |
| np.ones( N, dtype) | 生成一个N长度类型是dtype的一维全一ndarray |
| np.ones_like( ndarray ) | 生成一个形状与参数相同的全一ndarray |
| np.zeros( N) | 生成一个N长度的一维全零ndarray |
| np.zeros( N, dtype) | 生成一个N长度类型位dtype的一维全零ndarray |
| np.zeros_like(ndarray) | 类似np.ones_like( ndarray ) |
| np.empty( N ) | 生成一个N长度的未初始化一维ndarray |
| np.empty( N, dtype) | 生成一个N长度类型是dtype的未初始化一维ndarray |
| np.empty(ndarray) | 类似np.ones_like( ndarray ) |
| np.eye( N ) | 创建一个N * N的单位矩阵（对角线为1，其余为0） |
| np.identity( N ) | |
| np.arange( num) | 生成一个从0到num-1步数为1的一维ndarray |
| np.arange( begin, end) | 生成一个从begin到end-1步数为1的一维ndarray |
| np.arange( begin, end, step) | 生成一个从begin到end-step的步数为step的一维ndarray |
| np.mershgrid(ndarray, ndarray,...) | 生成一个ndarray * ndarray * ...的多维ndarray |
| np.where(cond, ndarray1, ndarray2) | 根据条件cond，选取ndarray1或者ndarray2，返回一个新的ndarray |
| np.in1d(ndarray, [x,y,...]) | 检查ndarray中的元素是否等于[x,y,...]中的一个，返回bool数组 |
import numpy as nparray 创建数组
np.array([[1,2],[3,4]])
arr = [1,2,3,4,5,6,7]
a = np.array(arr)
aarray([1, 2, 3, 4, 5, 6, 7])zeros 创建默认0的数组
a = np.zeros(3)
print ('a:',a)
b = np.zeros((3,3))
print('b:',b)
c = np.zeros_like(a)
print('c:',c)a: [0. 0. 0.]
b: [[0. 0. 0.]
[0. 0. 0.]
[0. 0. 0.]]
c: [0. 0. 0.]ones 创建默认1的数组
a = np.ones(3)
print ('a:',a)
b = np.ones((3,3))
print('b:',b)
c = np.ones_like(b)
print('c:',c)a: [1. 1. 1.]
b: [[1. 1. 1.]
[1. 1. 1.]
[1. 1. 1.]]
c: [[1. 1. 1.]
[1. 1. 1.]
[1. 1. 1.]]empty 创建默认值 随机的数组
print (np.empty((2)))
print(np.empty((2,4)))[-5.73021895e-300 6.92466535e-310]
[[6.92449667e-310 5.02034658e+175 6.03195894e+174 2.73950109e-057]
[7.12194759e-067 2.10979251e-052 1.47763641e+248 1.16096346e-028]]random.randn 随机值数组
np.random.randn(6)array([ 0.84344504, -0.73714262, 0.93808627, -1.00264675, -0.17058495,
0.4664123 ])random.randn 随机值多维数组
np.random.randn(2,2)array([[ 0.38153354, -0.92519611],
[-1.3137341 , -1.26921917]])arange 创建序列数组
## arange 0-9
print(np.arange(10))
## arange 5-15
print(np.arange(5,15))
## arange 13579
print(np.arange(1,10,2))[0 1 2 3 4 5 6 7 8 9]
[ 5 6 7 8 9 10 11 12 13 14]
[1 3 5 7 9]eye identity 创建对角线为一的多维数组
a = np.eye(4)
print(a)
np.identity(4)[[1. 0. 0. 0.]
[0. 1. 0. 0.]
[0. 0. 1. 0.]
[0. 0. 0. 1.]]
array([[1., 0., 0., 0.],
[0., 1., 0., 0.],
[0., 0., 1., 0.],
[0., 0., 0., 1.]])np.eye(3,3,3)array([[0., 0., 0.],
[0., 0., 0.],
[0., 0., 0.]])reshape 一维转多维
a = np.arange(4).reshape(2,2)
print(a)[[0 1]
[2 3]]meshgrid 生成一个ndarray * ndarray * ...的多维ndarray
x = np.array([0, 1, 2])
y = np.array([0, 1])
X, Y = np.meshgrid(x, y)
print(X)
print(Y)[[0 1 2]
[0 1 2]]
[[0 0 0]
[1 1 1]]常用方法
ndim 查看数据维度
shape 查看数据维度大小
dtype 查看数据类型
a = np.random.randn(3,3,3)
# 查看数据维度
print(a.ndim)
# 查看数据维度大小
print(a.shape)
# dtype 查看数据类型
print(a.dtype)3
(3, 3, 3)
float64astype转换数据类型
a = np.eye(5,5)
print(a.dtype)
b = a.astype(np.int64).dtype
print(b)float64
int64数学运算
# +
a = np.arange(4).reshape(2,2)
print(a)
print('++++++++')
print(a+1)
# -
print('--------')
print(a-1)
# *
print('********')
print(a*2)
# /
print('////////')
print(1/(a+1))[[0 1]
[2 3]]
++++++++
[[1 2]
[3 4]]
--------
[[-1 0]
[ 1 2]]
********
[[0 2]
[4 6]]
////////
[[1. 0.5 ]
[0.33333333 0.25 ]]比较运算
a = np.arange(6,0,-1).reshape(2,3)
b = np.arange(3,9).reshape(2,3)
print(a,b,a>b,sep='\n')[[6 5 4]
[3 2 1]]
[[3 4 5]
[6 7 8]]
[[ True True False]
[False False False]]切片和索引
一维切片同py list
a = np.arange(10)
print(a)
print(a[5])
print(a[5:8])
a[5:8] = 12
print(a)[0 1 2 3 4 5 6 7 8 9]
5
[5 6 7]
[ 0 1 2 3 4 12 12 12 8 9]切片后不等copy
b = a[5:8]
print(b)
b[1] = 21
print(a)[12 12 12]
[ 0 1 2 3 4 12 21 12 8 9]# 切片给数组赋值
a[:] = 0
aarray([0, 0, 0, 0, 0, 0, 0, 0, 0, 0])copy 复制
# 复制
b = a[2:5].copy()
print(b)
b[:] = 11
print(b)
print(a)[0 0 0]
[11 11 11]
[0 0 0 0 0 0 0 0 0 0]多维数组
# 多维数组
a2d = np.arange(6).reshape(2,3)
print(a2d)
# 两种取值方式,结果一样
print(a2d[0][2])
print(a2d[0,2])[[0 1 2]
[3 4 5]]
2
2# 多维数组
a3d = np.arange(1,13).reshape(2,2,3)
print(a3d)
print(a3d.ndim)
print(a3d[0])
print(a3d[0].ndim)[[[ 1 2 3]
[ 4 5 6]]
[[ 7 8 9]
[10 11 12]]]
3
[[1 2 3]
[4 5 6]]
2标量值和数组都可以赋值
# 标量值和数组都可以赋值给 a3d
tmp = a3d[0].copy()
a3d[0] = 99
print(a3d)
a3d[0] = tmp
print(a3d)[[[99 99 99]
[99 99 99]]
[[ 7 8 9]
[10 11 12]]]
[[[ 1 2 3]
[ 4 5 6]]
[[ 7 8 9]
[10 11 12]]]多维数组 切片索引
选行
a = np.eye(4)
# 选2 3 行
a[[2,3]]array([[0., 0., 1., 0.],
[0., 0., 0., 1.]])选列
a[:][2]array([0., 0., 1., 0.])a2d = np.arange(9).reshape(3,3)
print(a2d)
# 选择前2
print(a2d[:2])
# 选择 0 1 的0
print(a2d[:2,:1])
# 选择0 1 的 1 2
print(a2d[:2,1:])
# 选择 1 的 1 2
print(a2d[1,1:])
[[0 1 2]
[3 4 5]
[6 7 8]]
[[0 1 2]
[3 4 5]]
[[0]
[3]]
[[1 2]
[4 5]]
[4 5]a2d[:2,1:] = 0
a2darray([[0, 0, 0],
[3, 0, 0],
[6, 7, 8]])布尔型,索引
names = np.array(['Bob', 'Joe', 'Will', 'Bob', 'Will', 'Joe', 'Joe'])
data = np.random.randn(7,4)
print(names,data,sep='/n')
print(names == 'Bob')['Bob' 'Joe' 'Will' 'Bob' 'Will' 'Joe' 'Joe']/n[[-0.75689604 0.88583323 -1.49689044 0.95657837]
[-1.41937221 -0.22434237 1.74661417 0.08943732]
[ 1.55542971 -1.73335114 0.22718931 0.07517303]
[ 2.97463647 1.26849374 -0.38591486 0.1239141 ]
[-1.0057703 -0.26292255 1.42444036 -2.02698881]
[-0.00814679 0.38637696 -0.49233058 -1.13374044]
[-1.03599764 -1.3238696 -0.70148144 0.59657573]]
[ True False False True False False False]print(data[names == 'Bob'])
print(data[names == 'Bob',2:])
print(data[names == 'Bob',3])[[-0.75689604 0.88583323 -1.49689044 0.95657837]
[ 2.97463647 1.26849374 -0.38591486 0.1239141 ]]
[[-1.49689044 0.95657837]
[-0.38591486 0.1239141 ]]
[0.95657837 0.1239141 ]print(names != 'Bob')
print(data[~(names == 'Bob')])
print(((names == 'Bob') | (names == 'Will')))[False True True False True True True]
[[-1.41937221 -0.22434237 1.74661417 0.08943732]
[ 1.55542971 -1.73335114 0.22718931 0.07517303]
[-1.0057703 -0.26292255 1.42444036 -2.02698881]
[-0.00814679 0.38637696 -0.49233058 -1.13374044]
[-1.03599764 -1.3238696 -0.70148144 0.59657573]]
[ True False True True True False False]data[data<0] = 0
dataarray([[0. , 0.88583323, 0. , 0.95657837],
[0. , 0. , 1.74661417, 0.08943732],
[1.55542971, 0. , 0.22718931, 0.07517303],
[2.97463647, 1.26849374, 0. , 0.1239141 ],
[0. , 0. , 1.42444036, 0. ],
[0. , 0.38637696, 0. , 0. ],
[0. , 0. , 0. , 0.59657573]])data[names != 'Joe'] = 7
dataarray([[7. , 7. , 7. , 7. ],
[0. , 0. , 1.74661417, 0.08943732],
[7. , 7. , 7. , 7. ],
[7. , 7. , 7. , 7. ],
[7. , 7. , 7. , 7. ],
[0. , 0.38637696, 0. , 0. ],
[0. , 0. , 0. , 0.59657573]])a = np.empty((8,4))
for i in range(8):
a[i] = i
aarray([[0., 0., 0., 0.],
[1., 1., 1., 1.],
[2., 2., 2., 2.],
[3., 3., 3., 3.],
[4., 4., 4., 4.],
[5., 5., 5., 5.],
[6., 6., 6., 6.],
[7., 7., 7., 7.]])# 以特定顺序选取子集
print(a[[5,2,1]])
# 以顺序选取列
a[:,2][[5. 5. 5. 5.]
[2. 2. 2. 2.]
[1. 1. 1. 1.]]
array([0., 1., 2., 3., 4., 5., 6., 7.])# 负索引
print(a[[-3,-6,-7]])
print(a[[-3,-6,-7],[0,1,2]])
print(a[[-3,-6,-7]][:, [0,1]])[[5. 5. 5. 5.]
[2. 2. 2. 2.]
[1. 1. 1. 1.]]
[5. 2. 1.]
[[5. 5.]
[2. 2.]
[1. 1.]]数组转置和轴对换
a = np.arange(8).reshape((2,4))
print(a)
print(a.T)[[0 1 2 3]
[4 5 6 7]]
[[0 4]
[1 5]
[2 6]
[3 7]]dot计算内积
一维: x0*x1 + y0*y1
二维: arr2[0][0] = arr0[0][0] * arr1[0][0] + arr0[0][1] * arr1[1][0] (0*0 + 1*1)
arr2[0][1] = arr0[0][0] * arr1[0][1] + arr0[0][1] * arr1[1][1] (0*2 + 1*3)
arr2[1][0] = arr0[1][0] * arr1[0][0] + arr0[1][1] * arr1[1][0] (2*0 + 3*1)
arr2[1][1] = arr0[1][0] * arr1[0][1] + arr0[1][1] * arr1[1][1] (2*2 + 3*3)a = np.arange(4).reshape((2,2))
print(a)
print(a.T)
print(np.dot(a[0],a[1]))
print(np.dot(a, a.T))[[0 1]
[2 3]]
[[0 2]
[1 3]]
3
[[ 1 3]
[ 3 13]]transpose 交换坐标轴
维度 2 2 2 对应 012 相当交换维度
arr = np.arange(12).reshape((2, 2, 3))
print(arr)
arr.transpose((2,1,0)) [[[ 0 1 2]
[ 3 4 5]]
[[ 6 7 8]
[ 9 10 11]]]
array([[[ 0, 6],
[ 3, 9]],
[[ 1, 7],
[ 4, 10]],
[[ 2, 8],
[ 5, 11]]])arr = np.arange(9).reshape(3,3)
print(arr)
arr.transpose((1,0)) [[0 1 2]
[3 4 5]
[6 7 8]]
array([[0, 3, 6],
[1, 4, 7],
[2, 5, 8]])