numpy-就这些就够了

创建数组

import numpy as np
import math


# 创建数组对象 ndarray
a = np.arange(10)
print(a)
print(type(a))

# 对ndarray对象类型进行向量处理 sqrt 开平方
print(np.sqrt(a))

b = list(range(10))
c = []
for i in b:
    c.append(math.sqrt(i))
print(c)

import numpy as np


# 使用array函数创建一维数组1
a = np.array([1, 2, 3, 4])
print(a)
print(type(a))

# 使用array函数创建一维数组2
a1 = np.array([1, 2, 3, 4], dtype=float, ndmin=1)
print(a1)
print(type(a1))

# 使用array函数创建二维数组
b = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
print(b)
print(type(b))

# 使用array函数创建三维数组
c = np.array([[[1, 2, 3], [4, 5, 6], [7, 8, 9]]])
print(c)
print(type(c))

import numpy as np


# 使用arange创建数组
a = np.arange(1, 11)
print(a)

b = np.arange(1, 11, 2)
print(b)

print(type(a))

import numpy as np


# 使用random创建一维数组[0.0, 1.0)
a = np.random.random(size=5)
print(a)
print(type(a))

# 使用random创建二维数组[0.0, 1.0)
b = np.random.random(size=(3, 4))
print(b)
print(type(b))

# 使用random创建三维数组[0.0, 1.0)
c = np.random.random(size=(2, 3, 4)) # 两个3行4列
print(c)

# 随机整数生成一维数组
d = np.random.randint(5, 10, size=10)
print(d)

# 二维
e = np.random.randint(3, 9, size=(4, 3))
print(e)

# 三维
f = np.random.randint(9, size=(2, 4, 3))
print(d.dtype)
print(f)

import numpy as np


# 创建标准的正态分布 期望为0, 方差为1
# 一维
a = np.random.randn(4)
print(a)
print(type(a))

# 二维
b = np.random.randn(2, 3)
print(b)

# 三维
c = np.random.randn(2, 3, 4)
print(c)

# 创建指定期望值和方差的正态分布
d = np.random.normal(loc=4, scale=3, size=(2, 3, 4))
print(d)

import numpy as np


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

b = np.arange(6, 11)
c = np.random.random(5)
d = np.random.randint(100, 200, size=5)
e = np.random.randn(10)
f = np.random.normal(loc=3, scale=4, size=5)


# print(a)
# print(b)
# print(c)
# print(d)
# print(e)
# print(f)
print("ndim:", a.ndim, b.ndim, c.ndim, d.ndim)   # ndim 空间维数
print("shape:", a.shape, b.shape)                # shape 形状
print("dtype", a.dtype, b.dtype, c.dtype)
print("size", a.size, b.size, c.size, d.size, e.size, f.size)
print("itemsize:", a.itemsize, b.itemsize, c.itemsize)  # itemsize 每个元素所占的字节

import numpy as np

# 用0填充
a = np.zeros(5)
b = np.zeros((5, 5))
# 用1填充
c = np.ones(5)
# 创建空
d = np.empty((3, 3), dtype=int)
# np.linspace(start, stop, num=50, (endpoint=True  包括的), retstep=False, dtype=None)
# 该函数用于创建一个一维数组,数组是由一个等差数列构成的
e = np.linspace(1, 5, 10, endpoint=False)
# logspace (start, stop, num=50, endpoint, base, dtype) 创建等比数列
f = np.logspace(0, 9, 10, base=2)


print(a)
print(b)
print(c)
print(e)
print(d)
print(f)

一维数组切片

import numpy as np


# 一维数组的切片
a = np.arange(10)
print(a)
b = a[1::2]
print(b)
print(a[2])
print(a[-2])

二维数组切片

import numpy as np


a = np.arange(1, 13)
print(a)
# 一维数组转换成二维数组
x = a.reshape(4, 3)
print(x)
# 获取二维数组的第二行, 索引是从0开始的
print(x[1])
# 获取二维数组的第三行第二列
print(x[2][1])

# 切片的使用[行进行切片,列进行切片] [start:stop:step, start:stop:step]
# 获取所有行部分列
print(x[:, 1])
# 获取奇数行所有列
print(x[::2, :])
# 获取奇数行偶数列
print(x[::2, 1::2])

# 使用坐标获取数组[x, y]元素
print(x[2, 1])
# 同时获取第三行第二列,第四行第一列
print(x[2, 1], x[3, 0])
# 获取到的元素组成新的数组
print(np.array((x[2, 1], x[3, 0])))

# 负索引的使用
print("最后一行")
print(x[-1])
print("行倒序")
print(x[::-1])
print("行列倒序")
print(x[::-1, ::-1])

数组的复制, 深浅拷贝

import numpy as np


a = np.arange(1, 37).reshape(6, 6)
print(a)

b = a[0:2, 0:2]
print(b)

# 切片操作,对b的值进行修改,a的值也改变 是浅拷贝
b[0][0] = 100
print(b)
print(a)

# 深拷贝
c = np.copy(a[:2, :2])
print(c)

c[0][0] = 300
print(c)
print(b)

维度转换

import numpy as np


# 通过a.reshape变多维
a = np.arange(1, 37)
b = a.reshape(6, 6)
c = a.reshape(2, 3, 6)

# 通过np.reshape修改为多维
d = np.reshape(a, (2, 18))

print(a)
print(b)
print(c)
print(d)

# 1. 多变一
a1 = c.reshape(36)
# 2. 或者
a2 = c.reshape(-1)
print("3. 通过函数ravel")
a3 = c.ravel()
print("4. 通过flatten")
a4 = c.flatten()


print(a1)
print(a2)
print(a3)
print(a4)

数组的拼接

import numpy as np


a = np.arange(1, 13)
print(a)
a2 = a.reshape(4, 3)
print(a2)

b = np.arange(113, 125)
print(b)
b2 = b.reshape(4, 3)
print(b2)

print("水平拼接,用hstack, 注意一定要加[a2, b2], (a2, b2)也行")
c = np.hstack((a2, b2))
print(c)
print("垂直拼接vstract")
d = np.vstack([a2, b2])
print(d)
print("使用concatenate拼接,默认exis=0效果和vstract相同")
e = np.concatenate((a2, b2), axis=0)
print(e)
print("exis=1, 相当于hstract水平拼接")
f = np.concatenate((a2, b2), axis=1)
print(f)

print("三维数组有三个轴 axis= 0 1 2")
a3 = np.arange(1, 13).reshape(2, 2, 3)
b3 = np.arange(101, 113).reshape(2, 2, 3)
print(a3)
print(b3.shape)
c3 = np.concatenate((a3, b3), axis=0)
print(c3)
print(c3.shape)
d3 = np.concatenate((a3, b3), axis=1)
print(d3)
print(d3.shape)
e3 = np.concatenate((a3, b3), axis=2)
print(e3)
print(e3.shape)

数组的分割split

import numpy as np


print("一维数组的分割")
a = np.arange(1, 17)
print(a)
print(type(a))
b = np.split(a, 4)
print(b)
print(type(b))
print("像下面如果传递数组,就是分为3份")
c = np.split(a, [3, 5])
print(c)
print(type(c))
a2 = a.reshape(4, 4)
print(a2)
m = np.split(a2, 2)
print(m, type(m))
n = np.split(a2, 2, axis=1)
print(n)
print(type(n))
print("还有hsplit和vsplit")

数组转置

import numpy as np


print("数组的转置")
a = np.arange(1, 25).reshape(8, 3)
print(a)
print("transpose函数")
b = a.transpose()
print(b)
print("对于二维数组,也可以使用a.T转置")
c = a.T
print(c)
print("使用np.transpose()转置")
d = np.transpose(a)
print(d)
print("多维数组进行转置a[i][j][k],默认的将i和K交换 (下标2, 1, 0)")
a3 = np.arange(1, 25).reshape(2, 3, 4)
b3 = np.transpose(a3)
print(b3)
print(b3.shape)
print("多维数组转换")
c3 = np.transpose(a3, (1, 2, 0))
print(c3)
print(c3.shape)

数组的+,-,x,/, 三角函数

import numpy as np


print("数组的+, -, x, /")
print("加法")
a = np.arange(9).reshape(3, 3)
b = np.array([10, 10, 10])
print(a + b)
print(np.add(a, b))

print("减法")
print(b - a)
print(np.subtract(a, b))

print("乘法, out参数的使用")
y = np.empty((3, 3))
print(y)
np.multiply(a, 10, out=y)
print(y)

print("除法,同乘法")
np.divide(a, 10, out=y)
print(y)
print("三角函数")
a = np.array([0, 30, 60, 90])
print(np.sin(a * np.pi/180))

import numpy as np


a = np.array([0, 30, 60, 90])
print(np.sin(a * np.pi/180))

print("四舍五入函数, decimals: 舍去的小数位数,默认为0, 如果为负将四舍五入到小数点左侧的位置")
a1 = np.array([1.0, 4.45, 123, 0.657, 25.532])
aa = np.around(a1, decimals=1)
print(aa)
print("numpy.floor(), 返回数字的下舍整数")
a2 = np.floor(a1)
print(a2)
print("numpy.ceil(), 返回数字的上入整数")
a3 = np.ceil(a1)
print(a3)

print("统计函数:")
print("power", np.power(a, 2))
x = np.arange(5)
y = np.zeros(10)
np.power(2, x, out=y[:5])
print(y)

print("中位数:排完序后 偶数中间两个数的平均值,奇数是中间的数:median()", np.median(x))
print("二维数组 要通过axies指定轴, 0表示轴垂直, 1为水平方向, ", np.median(a, axis=0))
print("求平均值,mean", np.mean(a))
print("sum max min", np.max(a), np.sum(a), np.min(a))
print("argmin:最小值的下标,argmax: 最大值的下标", np.argmax(a), np.argmin(a))