Numpy模块(一)

Numpy是什么?

     它是一个开源科学计算库! 拥有丰富的数学函数,强大的多维数组以及优异的运算性能.

import numpy as np
import datetime as dt

n=100000
start=dt.datetime.now()
A,B = [] , []
for i in range(n):
    A.append( i**2)
    B.append (i**3)
C=[]
for a,b in zip(A,B):
    C.append(a+b)
print((dt.datetime.now()-start).microseconds)
start=dt.datetime.now()
# 170010

E=np.arange(n)**2+np.arange(n)**3
# E是一个数组
print((dt.datetime.now()-start).microseconds)
# 1000

numpy计算的时间能比python缩短100倍左右

多维数组:

 

    通过dtype参数和astype()方法都可以指定和修改元素的数据类型:

 

import numpy as np
import datetime as dt

a = np.array((10,23,32,23))
print(a)
# [10 23 32 23]
b = np.arange(1,10,2)
print(b)
# [1 3 5 7 9]
c = np.array([
    [1,2,3],
    [4,5,6]
])
print(c,type(c),type(c[0][0]))
print(c.dtype)
'''
 [[1 2 3]
 [4 5 6]] 
 <class 'numpy.ndarray'> <class 'numpy.int32'> 
 int32
'''
d = np.array(['1','2','3'])
print(type(d[0]),d.dtype)

# <class 'numpy.str_'>   <U1 (小端字节序,unicode编码,字符数为1)

e = np.array(['1','2','3'],dtype=np.int32)
print(e.dtype)
# int32
f = e.astype(np.str_)
print(f.dtype)
# <U11  整型转字符串填零变长了

 

import numpy as np
import datetime as dt

i = np.array([
    [np.arange(1,5),np.arange(5,9),np.arange(35,39)],
    [np.arange(13,17),np.arange(15,19),np.arange(45,49)]
])
print(i.shape)
# (2, 3, 4)  两页三行四页

循环打印多维数组中的每一个元素:

其中i,j,k分别表示页行列,a[i][j][k]和a[i,j,k]等价

 

示例:

方式一:

import numpy as np

a = np.array([('ABC',[1,2,3])],dtype = 'U3 , 3i4')

print(a, a[0]['f0'], a[0]['f1'])
# [('ABC', [1, 2, 3])]    ABC      [1 2 3]

方式二:

import numpy as np

b = np.array([('ABC',[1,2,3])],dtype = [
    ('name',np.str_,3),('score',np.int32,3)])

print(b[0]['name'] ,b[0]['score'])

# ABC [1 2 3]

方式三:

import numpy as np

c = np.array([('ABC',[1, 2, 3])],dtype = {
    'names':['name','score'],
    'formats':['U3' , '3i4']})
print(c)

print(c[0]['name'] ,c[0]['score'])

# [('ABC', [1, 2, 3])]    ABC    [1 2 3]

方式四:

import numpy as np

d = np.array([('ABC',[1, 2, 3])],dtype = {
    'name':('U3',0),
    'score':('3i4',12) })
print(d)

print(d[0]['name'] ,d[0]['score'])

# [('ABC', [1, 2, 3])]    ABC    [1 2 3]

 方式五:

import numpy as np

e = np.array([0x1234],dtype=(
    '<u2',{'lo':('u1',0),'hi':('u1',1)}))
print('{:x} {:x}'.format(e['lo'][0], e['hi'][0]))
# 小端字节序,高字段低地址 34 12

e = np.array([0x1234],dtype=(
    '>u2',{'lo':('u1',0),'hi':('u1',1)}))
print('{:x} {:x}'.format(e['lo'][0], e['hi'][0]))
# 大端字节序,低字段高地址 12 34

一维数组的切分:

import numpy as np

a = np.arange(1,10)
print((a[:3]))    # 1 2 3
print(a[3:6])     # 4 5 6
print(a[6:])    # 7 8 9
print(a[::-1])  #  9 8 7 6 5 4 3 2 1
print(a[:-4:-1])  # 9 8 7
print(a[-4:-7:-1]) # 6 5 4
print(a[...]) # 1 2 3 4 5 6 7 8 9
print(a[::]) # 1 2 3 4 5 6 7 8 9
print(a[:]) # 1 2 3 4 5 6 7 8 9

多维数组的切分:

b = np.arange(1,25).reshape(2,3,4)
print(b)
'''
[[[ 1  2  3  4]
  [ 5  6  7  8]
  [ 9 10 11 12]]

 [[13 14 15 16]
  [17 18 19 20]
  [21 22 23 24]]]
'''
print(b[:,0,0])
#  所有页的第0行第0列的数据  [ 1 13]
print(b[0,:,:])
'''
第0页所有数据
[[ 1  2  3  4]
 [ 5  6  7  8]
 [ 9 10 11 12]]
'''
print(b[0,1,::2])
#第0页第1行 索引从0步长为2 拿所有数据 [5 7]

print(b[:,:,1])
print(b[...,1])
'''
所有页所有行的第1列
[[ 2  6 10]
 [14 18 22]]
'''
print(b[:,1])
'''
所有页的第1行的数据
[[ 5  6  7  8]
 [17 18 19 20]]
'''
print(b[-1,1:,2:])
'''
最后一页第一行及以后第二列及以后的所有数据
[[19 20]
 [23 24]]

视图与副本示例:

import numpy as np

a = np.arange(1,9)

b = a.reshape(2,4)
c = a.reshape(2,2,2)
print(b)
print(c)
'''
[[1 2 3 4]
 [5 6 7 8]]

[[[1 2]
  [3 4]]
[[5 6]
  [7 8]]]

'''
d = c.ravel()
e = c.flatten()
print(d)
print(e)
'''
ravel    一维视图,数据仍是源数据
flatten  一维副本,数据为自己数据

[1 2 3 4 5 6 7 8]
[1 2 3 4 5 6 7 8]

'''

a += 10
print(d)
print(e)
'''
[11 12 13 14 15 16 17 18]
[1 2 3 4 5 6 7 8]
'''
补充:
多维副本:
f = b.reshape(2,2,2).copy()

一维数组转置两种处理方式:

import numpy as np

a = np.arange(1,10).reshape(3,3)
b = np.arange(11,20).reshape(3,3)
c = np.vstack((a,b))
print(c)
'''
垂直堆叠
[[ 1  2  3]
 [ 4  5  6]
 [ 7  8  9]
 [11 12 13]
 [14 15 16]
 [17 18 19]]


'''


e,f = np.vsplit(c,2)
print(e,f,sep='\n')
'''
垂直拆分为2个相同维度的数组
[[1 2 3]
 [4 5 6]
 [7 8 9]]

[[11 12 13]
 [14 15 16]
 [17 18 19]]

'''

import numpy as np

a = np.arange(1,10).reshape(3,3)
b = np.arange(11,20).reshape(3,3)
c = np.hstack((a,b))
print(c)
'''
水平堆叠
[[ 1  2  3 11 12 13]
 [ 4  5  6 14 15 16]
 [ 7  8  9 17 18 19]]


'''
e,f = np.hsplit(c,2)
print(e,f,sep='\n')
'''
水平拆分为2个相同维度的数组
[[1 2 3]
 [4 5 6]
 [7 8 9]]

[[11 12 13]
 [14 15 16]
 [17 18 19]]

'''

 

import numpy as np

a = np.arange(1,10).reshape(3,3)
b = np.arange(11,20).reshape(3,3)
c = np.dstack((a,b))
print(a,b,sep='\n')
print(c)
'''
[[1 2 3]
 [4 5 6]
 [7 8 9]]

[[11 12 13]
 [14 15 16]
 [17 18 19]]

前后每行元素一一对应
[[[ 1 11]
  [ 2 12]
  [ 3 13]]

 [[ 4 14]
  [ 5 15]
  [ 6 16]]

 [[ 7 17]
  [ 8 18]
  [ 9 19]]]


'''
e,f = np.dsplit(c,2)
print(e,f,sep='\n')
'''
拆分后不能恢复原来数组
[[[1]
  [2]
  [3]]

 [[4]
  [5]
  [6]]

 [[7]
  [8]
  [9]]]

[[[11]
  [12]
  [13]]

 [[14]
  [15]
  [16]]

 [[17]
  [18]
  [19]]]

'''

print(e.T[0].T,f.T[0].T,sep='\n')
'''
转置操作后能恢复为原来数组:
[[1 2 3]
 [4 5 6]
 [7 8 9]]
[[11 12 13]
 [14 15 16]
 [17 18 19]]
'''

import numpy as np

a = np.arange(1,10)
b = np.arange(11,20)
c = np.row_stack((a,b))
print(c)
'''
与vstack处理的效果相同
[[ 1  2  3  4  5  6  7  8  9]
 [11 12 13 14 15 16 17 18 19]]

'''
d = np.column_stack((a,b))
print(d)
'''
与hstack效果不同
[[ 1 11]
 [ 2 12]
 [ 3 13]
 [ 4 14]
 [ 5 15]
 [ 6 16]
 [ 7 17]
 [ 8 18]
 [ 9 19]]
'''

 

import numpy as np

a = np.array([
    [1+1j, 2+2j, 3+3j],
    [4+4j, 5+5j, 6+6j],
    [7+7j, 8+8j, 9+9j]
])
print(a.dtype,a.dtype.str,a.dtype.char)
# complex128 <c16 D  每个元素16的字节,128位二进制数
print(a.shape)
#(3, 3) 三行三列
print(a.ndim)
#每个元素是二维的
print(a.size,len(a))
#size指元素个数为9,len指数组行数3
print(a.itemsize,a.nbytes)
#16 144  每个元素16的字节,总字节数为9*16=144
print(a.T)
'''
转置
[[1.+1.j 4.+4.j 7.+7.j]
 [2.+2.j 5.+5.j 8.+8.j]
 [3.+3.j 6.+6.j 9.+9.j]]
'''
print(a.real,a.imag,sep='\n')
'''
实部,虚部各自打印
[[1. 2. 3.]
 [4. 5. 6.]
 [7. 8. 9.]]

[[1. 2. 3.]
 [4. 5. 6.]
 [7. 8. 9.]]
'''
for elem in a.flat:
    print(elem)
'''
迭代取值
(1+1j)
(2+2j)
(3+3j)
(4+4j)
(5+5j)
(6+6j)
(7+7j)
(8+8j)
(9+9j)
'''
print(a.flat[[1,3,5]])
a.flat[[2,4,6]] = 0
print(a)
'''
flat迭代器可按索引取值
[2.+2.j 4.+4.j 6.+6.j]

也可以修改值
[[1.+1.j 2.+2.j 0.+0.j]
 [4.+4.j 0.+0.j 6.+6.j]
 [0.+0.j 8.+8.j 9.+9.j]]
'''

数组与列表的区别与转换:

import numpy as np

def fun(a,b):
    a.append(b)
    return a

x = np.array([10,20,30])
y = 40
x = np.array(fun(x.tolist(),y))
print(x)
# [10 20 30 40] 数组.tolist()转换成列表就可以append了
x=np.append(x,50)
print(x)
# [10 20 30 40 50] 数组可用append函数添加数据,但是必须用x接受方可打印