高强度快速入门numpy，节选自官方manual

最新推荐文章于 2026-07-02 16:59:22 发布

原创最新推荐文章于 2026-07-02 16:59:22 发布 · 535 阅读

0 ·

本内容遵循CC 4.0 BY-SA版权协议

收录于

python库基础

本文详细介绍了NumPy库的基础使用，包括数组创建、基本运算、通用函数、索引与切片等，并深入探讨了数组形状操作、复制与视图、广播规则及高级索引技巧，适合初学者快速上手并深入掌握。

起因

是我很长时间没有用python的numpy了，于是重新啃了啃原始官方manual。
版本：numpy v1.19
点击这里从官网下载：https://numpy.org/doc/1.19/numpy-user.pdf

import numpy as np

2.2.2 Array Creation

按照python列表创建

创建全0，全1的array

创建一个空的array

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

array([[1, 2, 3],
       [3, 4, 5]])

np.zeros((3,4))

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

此外还有zeros_like函数

np.ones( (2,3,4), dtype=np.int16 )

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]]], dtype=int16)

此外还有ones_like函数

np.empty( (2,3) )

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

此外还有empty_like函数

创建一些列值的数组：

arange(开始值，结束值，步长)，和python的ranage用法是一样的
linspace(开始值，结束值，划分数量）

np.arange( 10, 30, 5 )

array([10, 15, 20, 25])

np.arange( 0, 2, 0.3 )

array([0. , 0.3, 0.6, 0.9, 1.2, 1.5, 1.8])

np.linspace( 0, 2, 9 )

array([0.  , 0.25, 0.5 , 0.75, 1.  , 1.25, 1.5 , 1.75, 2.  ])

x = np.linspace( 0, 2*np.pi, 22 )    ##可以处理比较复杂的序列划分
f = np.sin(x)
print(f)

[ 0.00000000e+00  2.94755174e-01  5.63320058e-01  7.81831482e-01
  9.30873749e-01  9.97203797e-01  9.74927912e-01  8.66025404e-01
  6.80172738e-01  4.33883739e-01  1.49042266e-01 -1.49042266e-01
 -4.33883739e-01 -6.80172738e-01 -8.66025404e-01 -9.74927912e-01
 -9.97203797e-01 -9.30873749e-01 -7.81831482e-01 -5.63320058e-01
 -2.94755174e-01 -2.44929360e-16]

2.2.3 Printing Arrays

# np.set_printoptions(threshold=sys.maxsize)## 通过这个指令可以让输出的时候不再有很多省略，全部输出

2.2.4 Basic Operations

两个array的乘法：点乘和叉乘

A = np.array( [[1,1],[0,1]] )
B = np.array( [[2,0],[3,4]] )

A * B # elementwise product

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

A @ B # matrix product

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

A.dot(B) # another matrix product

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

有一些和数组无关的操作也会直接集成在ndarray类中

rg = np.random.default_rng(1)
a224 = rg.random((2,3))
print(a224)
print("sum: ",a224.sum())
print("min: ",a224.min())
print("max: ",a224.max())

[[0.51182162 0.9504637  0.14415961]
 [0.94864945 0.31183145 0.42332645]]
sum:  3.290252281866131
min:  0.14415961271963373
max:  0.9504636963259353

2.2.5 Universal Functions

包含了一些常见的操作与数学运算，比如exp, sqrt, add, all, any, apply_along_axis, argmax, argmin, argsort, average, bincount, ceil, clip, conj,
corrcoef, cov, cross, cumprod, cumsum, diff, dot, floor, inner, invert, lexsort, max, maximum,
mean, median, min, minimum, nonzero, outer, prod, re, round, sort, std, sum, trace, transpose,
var, vdot, vectorize, where 等等

2.2.6 Indexing, Slicing and Iterating

一维的数组可以被索引，划分和迭代。和python中的列表差不多

a = np.arange(10)**3
print(a)
a[::-1]    ##列表翻转小技巧

[  0   1   8  27  64 125 216 343 512 729]





array([729, 512, 343, 216, 125,  64,  27,   8,   1,   0], dtype=int32)

多维数组每一个axis有一个index，是一个用逗号分隔的元组

b = np.fromfunction(lambda x,y:10*x+y,(5,4), dtype=int)
b

array([[ 0,  1,  2,  3],
       [10, 11, 12, 13],
       [20, 21, 22, 23],
       [30, 31, 32, 33],
       [40, 41, 42, 43]])

b[2,3]

b[0:5, 1]     # each row in the second column of b

array([ 1, 11, 21, 31, 41])

b[ : ,1] # equivalent to the previous example

array([ 1, 11, 21, 31, 41])

b[1:3, : ] # each column in the second and third row of b

array([[10, 11, 12, 13],
       [20, 21, 22, 23]])

当拥有小于维度的参数时，后面默认为:

b[-1] # the last row. Equivalent to b[-1,:]

array([40, 41, 42, 43])

“…”这样的表示方式是有尽可能多的列，也就是尽可能多的“:”

The dots (…) represent as many colons as needed to produce a complete indexing tuple

x[1,2,…] is equivalent to x[1,2,:,:,:]
x[…,3] to x[:,:,:,:,3] and
x[4,…,5,:] to x[4,:,:,5,:].

c = np.array( [[[ 0, 1, 2], [ 10, 12, 13]],[[100,101,102],[110,112,113]]])
print(c.shape)
print(c[1,...])
print(c[...,2])

(2, 2, 3)
[[100 101 102]
 [110 112 113]]
[[  2  13]
 [102 113]]

如果需要挨个打印元素的话，可以使用flat属性

for elem in c.flat:
    print(elem)

2.3 Shape Manipulation

2.3.1 Changing the shape of an array

a = np.floor(10*rg.random((3,4)))
print(a)
print("shape: ",str(a.shape))

[[8. 4. 5. 0.]
 [7. 5. 3. 7.]
 [3. 4. 1. 4.]]
shape:  (3, 4)

下面的三个指令返回一个修改过的array，但是不会修改原array

ravel()
reshape()
T

a.ravel() # returns the array, flattened

array([8., 4., 5., 0., 7., 5., 3., 7., 3., 4., 1., 4.])

a.reshape(6,2) # returns the array with a modified shape

array([[8., 4.],
       [5., 0.],
       [7., 5.],
       [3., 7.],
       [3., 4.],
       [1., 4.]])

a.T # returns the array, transposed

array([[8., 7., 3.],
       [4., 5., 4.],
       [5., 3., 1.],
       [0., 7., 4.]])

a.T.shape

(4, 3)

a.shape  #并没有修改原始数据

(3, 4)

reshape函数会创建一个array的副本，而resize函数会改变array内容本身

如果维度参数是-1，那么这个参数就会被自动计算

print(a)
a.resize((2,6))
print(a)

[[8. 4. 5. 0.]
 [7. 5. 3. 7.]
 [3. 4. 1. 4.]]
[[8. 4. 5. 0. 7. 5.]
 [3. 7. 3. 4. 1. 4.]]

a = a.reshape(3,-1)
print(a)

[[8. 4. 5. 0.]
 [7. 5. 3. 7.]
 [3. 4. 1. 4.]]

2.3.2 Stacking together different arrays

a = np.floor(10*rg.random((2,2)))
print("a: ",a)
b = np.floor(10*rg.random((2,2)))
print("b: ",b)
print("vstack: ",np.vstack((a,b)))
print("hstack: ",np.hstack((a,b)))

a:  [[5. 2.]
 [1. 9.]]
b:  [[5. 1.]
 [6. 7.]]
vstack:  [[5. 2.]
 [1. 9.]
 [5. 1.]
 [6. 7.]]
hstack:  [[5. 2. 5. 1.]
 [1. 9. 6. 7.]]

函数column_stack将一维的数组作为column叠到二维数组中,此外还有row_stack操作

from numpy import newaxis
np.column_stack((a,b))

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

a = np.array([4.,2.])
b = np.array([3.,8.])
print(np.column_stack((a,b))) # returns a 2D array
print(np.hstack((a,b)))  # the result is different

[[4. 3.]
 [2. 8.]]
[4. 2. 3. 8.]

print(a[:,newaxis])  # view `a` as a 2D column vector
print(np.column_stack((a[:,newaxis],b[:,newaxis])))
print(np.hstack((a[:,newaxis],b[:,newaxis])))  # the result is the same

[[4.]
 [2.]]
[[4. 3.]
 [2. 8.]]
[[4. 3.]
 [2. 8.]]

2.3.3 Splitting one array into several smaller ones

a = np.floor(10*rg.random((2,12)))
a

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

np.hsplit(a,3)   #Split a into 3

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

np.hsplit(a,(3,4))  # Split a after the third and the fourth column

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

2.4 Copies and Views

2.4.1 No Copy at All

b=a

2.4.2 View or Shallow Copy

c = a.view()

2.4.3 Deep Copy

d = a.copy()

2.4.4 Functions and Methods Overview

Array Creation

arange, array, copy, empty, empty_like, eye, fromfile, fromfunction, identity, linspace, logspace, mgrid, ogrid, ones, ones_like, r_, zeros, zeros_like

Conversions

ndarray.astype, atleast_1d, atleast_2d, atleast_3d, mat Manipulations
array_split, column_stack, concatenate, diagonal, dsplit, dstack, hsplit, hstack, ndarray.item, newaxis, ravel, repeat, reshape, resize, squeeze, swapaxes, take, transpose, vsplit, vstack

Questions

all, any, nonzero, where

Ordering

argmax, argmin, argsort, max, min, ptp, searchsorted, sort

Operations

choose, compress, cumprod, cumsum, inner, ndarray.fill, imag, prod, put, putmask, real,sum

Basic Statistics

cov, mean, std, var

Basic Linear Algebra

cross, dot, outer, linalg.svd, vdot

2.5 Less Basic(一些高级内容：广播)

2.5.1 Broadcasting rules

2.6 Advanced indexing and index tricks

2.6.1 Indexing with Arrays of Indices

a = np.arange(12)**2 # the first 12 square numbers
i = np.array([1, 1, 3, 8, 5]) # an array of indices
a[i] # the elements of a at the positions i

array([ 1,  1,  9, 64, 25], dtype=int32)

j = np.array([[3, 4], [9, 7]]) # a bidimensional array of indices
a[j] # the same shape as j

array([[ 9, 16],
       [81, 49]], dtype=int32)

palette = np.array([[0, 0, 0], # black
    [255, 0, 0],               # red
    [0, 255, 0],               # green
    [0, 0, 255],               # blue
    [255, 255, 255]])          # white
image = np.array([[0, 1, 2, 0], # each value corresponds to a color in the palette
                [0, 3, 4, 0]])
palette[image] # the (2, 4, 3) color image

array([[[  0,   0,   0],
        [255,   0,   0],
        [  0, 255,   0],
        [  0,   0,   0]],

       [[  0,   0,   0],
        [  0,   0, 255],
        [255, 255, 255],
        [  0,   0,   0]]])

后面还有一些高级的索引方式和元素操作，我看不懂就不写了