Numpy&Pandas(byMorvan)-numpy

一、Numpy

1 概述

1.1 numpy&pandas有什么用

1.2 numpy&pandas安装

pip install numpy 
pip install pandas

2 基本操作

2.1 numpy 属性

import numpy as np

array = np.array([[1, 2, 3], [2, 3, 4]])  # 列表转换为数组

print(array)  
'''
[[1 2 3] 
[2 3 4]]
''' 

print("形状", array.shape)  # 形状 (2,3)
print("维度", array.ndim)  # 维度 2
print("元素个数", array.size)  # 元素个数 6

2.2 numpy 创建array

import numpy as np

# 生成一维数组
a = np.array([[2, 23, 4]], dtype=np.int32)  # dtype定义类型，默认是32位
print(a, a.dtype)  # [ 2 23  4]

# 生成二维数组
a = np.array([[2, 23, 4], [2, 32, 4]])
print(a)
'''
[[ 2 23  4]
 [ 2 32  4]]
'''

# 生成全部为0的矩阵
a = np.zeros((3, 4))  # 生成全部为0，3行4列
print(a)
'''
[[0. 0. 0. 0.]
 [0. 0. 0. 0.]
 [0. 0. 0. 0.]]
'''

# 生成全部为1的矩阵
a = np.ones((3, 4), dtype=np.int16)  # 生成全部为1
print(a)
'''
[[1. 1. 1. 1.]
 [1. 1. 1. 1.]
 [1. 1. 1. 1.]]
'''

# 生成空矩阵
a = np.empty((3, 4))
print(a)
'''
[[0. 0. 0. 0. 0.]
 [0. 0. 0. 0. 0.]
 [0. 0. 0. 0. 0.]]
'''

# 生成有序矩阵，arange(开始,结束,间隔)
a = np.arange(10, 20, 2)
print(a)  # [10 12 14 16 18]

# 生成有序矩阵，并重新定义形状
a = np.arange(12).reshape((3, 4))  # 生成有序数列,重新定义形状
print(a)
'''
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
'''

# 生成线段矩阵，linspace(开始,结束,段数)
a = np.linspace(1, 10, 6).reshape((2, 3))  # 生成线段,在1和10之间生成6段数列，默认50
print(a)
'''
[[ 1.   2.8  4.6]
 [ 6.4  8.2 10. ]]
'''

2.3 numpy 基础运算1

import numpy as np

a = np.array([10, 20, 30, 40])
b = np.arange(4)
print(a, b)  # [10 20 30 40] [0 1 2 3]
c = a + b
print(c)  # [10 21 32 43]
c = a - b
print(c)  # [10 19 28 37]
b = b ** 2
print(b)  # [0 1 4 9]

c = np.sin(a)
print(c)  # [-0.54402111  0.91294525 -0.98803162  0.74511316]

print(b > 4)  # [False False False  True]

c = b[b > 2]
print(c)  # [4 9]

a = np.array([[1, 1], [2, 2]])
b = np.arange(4).reshape((2, 2))
print(a, "\n", b)
'''
[[1 1]
 [2 2]] 
 [[0 1]
 [2 3]]
'''

c = a * b
print(c)  # 逐个相乘
'''
[[0 1]
 [4 6]]
'''

c = np.dot(a, b)  # 矩阵乘法
print(c)
'''
[[2 4]
 [4 8]]
'''

c = a.dot(b)
print(c)
'''
[[2 4]
 [4 8]]
'''

# 随机生成
a = np.random.random((2, 4))
print(a)
'''
[[0.28505404 0.73788789 0.92095046 0.01643222]
 [0.86400751 0.22622767 0.85907991 0.32050644]]
'''

print(np.sum(a))  # 求和
print(np.min(a))  # 最大值
print(np.max(a))  # 最小值

print(np.sum(a, axis=1))  # axis=0对列运算，axis=1对行运算
'''
[1.92030221 1.66420099]
'''

print(np.min(a, axis=1))  # axis=0对列运算，axis=1对行运算
'''
[0.18022666 0.10125785]
'''

print(np.max(a, axis=0))  # axis=0对列运算，axis=1对行运算
'''
[0.82491043 0.33942129 0.6142982  0.95921991]
'''

2.4 numpy 基础运算2

import numpy as np

A = np.arange(2, 14).reshape((3, 4))
print(A)  # 获取索引
'''
[[ 2  3  4  5]
 [ 6  7  8  9]
 [10 11 12 13]]
'''

print(np.argmin(A))  # 获取最小值的索引 0
print(np.argmax(A))  # 获取最大值索引 11

print(np.mean(A))  # 平均值 7.5
print(A.mean())  # 平均值 7.5
print(np.average(A))  # 平均值 7.5
print(np.median(A))  # 中位数 7.5

print(np.cumsum(A))  # 逐步求和 [ 2  5  9 14 20 27 35 44 54 65 77 90]
print(np.diff(A))  # 逐步求差 [[1 1 1] [1 1 1] [1 1 1]]

print(np.nonzero(A))  # 输出非零的行列数目，两个array，分别代表行数和列数
''' 
(array([0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2], dtype=int64),array([0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3], dtype=int64))
'''

print(np.sort(A))  # 逐行排序

print(np.transpose(A))  # 矩阵转置
'''
[[ 2  6 10]
 [ 3  7 11]
 [ 4  8 12]
 [ 5  9 13]]
'''

print(A.T.dot(A))  # 矩阵乘法(转置后)
'''
[[140 158 176 194]
 [158 179 200 221]
 [176 200 224 248]
 [194 221 248 275]]
'''

print(np.clip(A, 5, 9))  # 大于9变成9，小于5变成5
'''
[[5 5 5 5]
 [6 7 8 9]
 [9 9 9 9]]
'''

print(np.mean(A, axis=0))  # 列平均 [6. 7. 8. 9.]
print(np.mean(A, axis=1))  # 行平均 [ 3.5  7.5 11.5]

2.5 numpy 索引

import numpy as np

A = np.arange(3, 15).reshape((3, 4))
print(A)
'''
[[ 3  4  5  6]
 [ 7  8  9 10]
 [11 12 13 14]]
'''

print(A[2])  # 输出第一行 [11 12 13 14]
print(A[2][1])  # 输出第一行，第一列 12
print(A[2, 1])  # 输出第一行，第一列 12

print(A[2, :])  # 输出第二行所有数 [11 12 13 14]
print(A[:, 1])  # 输出第一列所有数 [ 4  8 12]
print(A[1, 1:3])  # 输出第一行第二列的数，左闭右开 [8 9]

for row in A:  # 迭代行
    print(row)
'''
[3 4 5 6]
[ 7  8  9 10]
[11 12 13 14]
'''

for column in A.T:  # 迭代列
    print(column)
'''
[ 3  7 11]
[ 4  8 12]
[ 5  9 13]
[ 6 10 14]
'''

print(A.flatten())  # [ 3  4  5  6  7  8  9 10 11 12 13 14]
for item in A.flat:  # 迭代每一项
    print(item, end="\t")  # 3	4	5	6	7	8	9	10	11	12	13	14

2.6 numpy array合并

import numpy as np

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

C = np.vstack((A, B))  # 上下合并
print(C, C.shape)
'''
[[1 1 1]
 [2 2 2]] (2, 3)
'''

D = np.hstack((A, B))  # 左右合并
print(D, D.shape)  # [1 1 1 2 2 2] (6,)

print(A.shape, A[np.newaxis, :].shape, A[:, np.newaxis].shape)  # (3,) (1, 3) (3, 1)
print(A, A[np.newaxis, :], A[:, np.newaxis])  # 行新加一个维度  [1 1 1] [[1 1 1]]
print(A[:, np.newaxis])  # 行新加一个维度  [1 1 1] [[1 1 1]]
'''
[[1]
 [1]
 [1]]
'''
A = A[:, np.newaxis]
B = B[:, np.newaxis]
D = np.hstack((A, B))  # 左右合并
print(D)
'''
[[1 2]
 [1 2]
 [1 2]]
'''

C = np.concatenate((A, B, B, A), axis=1)  # 0列 1行
print(C)
'''
axis=0
[[1]
 [1]
 [1]
 [2]
 [2]
 [2]
 [2]
 [2]
 [2]
 [1]
 [1]
 [1]]
axis=1
[[1 2 2 1]
 [1 2 2 1]
 [1 2 2 1]]
'''

2.7 numpy array分割

import numpy as np

A = np.arange(12).reshape((3, 4))
print(A)

# 等量分割
print(np.split(A, 2, axis=1))  # 1表示把行分成两部分
'''
[array([[0, 1],
       [4, 5],
       [8, 9]]), array([[ 2,  3],
       [ 6,  7],
       [10, 11]])]
'''
print(np.hsplit(A, 2))  # 水平分割
'''
[array([[0, 1],
       [4, 5],
       [8, 9]]), array([[ 2,  3],
       [ 6,  7],
       [10, 11]])]
'''

print(np.split(A, 3, axis=0))  # 0表示把列分成三部分 [array([[0, 1, 2, 3]]), array([[4, 5, 6, 7]]), array([[ 8,  9, 10, 11]])]
print(np.vsplit(A, 3))  # 纵向分割 [array([[0, 1, 2, 3]]), array([[4, 5, 6, 7]]), array([[ 8,  9, 10, 11]])]

# 不等量分割
print(np.array_split(A, 3, axis=1))
'''
[array([[0, 1],
       [4, 5],
       [8, 9]]), array([[ 2],
       [ 6],
       [10]]), array([[ 3],
       [ 7],
       [11]])]
'''

2.8 numpy copy&deep copy

import numpy as np

a = np.arange(4)
b = a
c = a
d = b
a[0] = 9
print(a, b, c, d)  # [9 1 2 3] [9 1 2 3] [9 1 2 3] [9 1 2 3]

b = a.copy()  # deep copy
print(b)
a[3] = 44
print(a, b)  # [ 9  1  2 44] [9 1 2 3]