数组的创建
import numpy as np
t1 = np.array([1,2,3])
print(t1,type(t1))
>>>[1 2 3] <class 'numpy.ndarray'>
t2 = np.array(range(10))
print(t2,type(t2))
>>>[0 1 2 3 4 5 6 7 8 9] <class 'numpy.ndarray'>
t3 = np.arange(10)#跟range用法一样，np.arrage(1,10,2)
print(t3,type(t3))
>>>[0 1 2 3 4 5 6 7 8 9] <class 'numpy.ndarray'>
numpy中常见的数据类型：
查看数据类型：
t1 = np.array(range(1,6))
print(t1)
print(t1.dtype)
>>>[1 2 3 4 5]
>>>int32
指定数据类型：
t1 = np.array(range(1,6),dtype=float) #方式二：dtype="int8" #方式三：dtype="i1"
print(t1)
print(t1.dtype)
>>>[1. 2. 3. 4. 5.]
>>>float64
#布尔类型
t1 = np.array([1,0,1,1],dtype=bool)
print(t1)
print(t1.dtype)
>>>[ True False True True]
>>>bool
修改数据类型：
t1 = np.array([1,0,1,1],dtype=bool)
print(t1)
print(t1.dtype)
>>>[ True False True True]
>>>bool
t2 = t1.astype("int8")
print(t2)
print(t2.dtype)
>>>[1 0 1 1]
>>>int8
修改浮点型保留小数位数：
t1 = np.array([random.random() for i in range(3)])
print(t1)
print(t1.dtype)
>>>[0.63296597 0.04263672 0.92288572]
>>>float64
t2 = np.round(t1,2)
print(t2)
>>>[0.63 0.04 0.92]
数组的形状
#一维数组
t1 = np.arange(5)
print(t1)
print(t1.shape)
>>>[0 1 2 3 4]
>>>(5,)
t2 = np.array([[1,2,3],[4,5,6]])
print(t2)
print(t2.shape)
>>>[[1 2 3]
[4 5 6]]
>>>(2, 3)
修改数组形状，reshape()方法不会改变数组本身，会返回一个新的数组。reshape((12,1))或者reshape(12,1)都行，官方文档使用的是元祖
#一维数组
t1 = np.arange(12)
print(t1)
print(t1.shape)
>>>[ 0 1 2 3 4 5 6 7 8 9 10 11]
>>>(12,)
#t1修改成二维数组，12行1列
t2 = t1.reshape(12,1)
print(t2)
>>>
[[ 0]
[ 1]
[ 2]
[ 3]
[ 4]
[ 5]
[ 6]
[ 7]
[ 8]
[ 9]
[10]
[11]]
#t1修改成二维数组，1行12列
t3 = t1.reshape(1,12)
print(t3)
>>>[[ 0 1 2 3 4 5 6 7 8 9 10 11]]
#t1修改成二维数组3行4列
t4 = t1.reshape((3,4))
print(t4)
>>>[[ 0 1 2 3]
[ 4 5 6 7]
[ 8 9 10 11]]
#t1修改成二维数组3行5列报错
t5 = t1.reshape((3,5))
print(t5)
>>>Traceback (most recent call last):
File "test.py", line 12, in <module>
t3 = t1.reshape((3,5))
ValueError: cannot reshape array of size 12 into shape (3,5)
#t4修改成一维数组
t6 = t4.reshape((t4.shape[0]*t4.shape[1],))
print(t6)
>>>[ 0 1 2 3 4 5 6 7 8 9 10 11]
#t4修改成一位数组，简便方法
t7 = t4.flatten()
print(t7)
>>>[ 0 1 2 3 4 5 6 7 8 9 10 11]
数组的计算
t1 = np.arange(12).reshape(3,4)
print(t1)
>>>
[[ 0 1 2 3]
[ 4 5 6 7]
[ 8 9 10 11]]
#每个元素都+10
t2 = t1+10
print(t2)
>>>
[[10 11 12 13]
[14 15 16 17]
[18 19 20 21]]
#每个元素都除以0，0理解成非常小的数
t3 = t1/0
print(t3)
>>>
[[nan inf inf inf]
[inf inf inf inf]
[inf inf inf inf]]
#相同形状的数组相加，即对应位置的元素相加
t4 = np.arange(12,24).reshape(3,4)
print(t4)
>>>
[[12 13 14 15]
[16 17 18 19]
[20 21 22 23]]
t5 = t1 + t4
print(t5)
>>>
[[12 14 16 18]
[20 22 24 26]
[28 30 32 34]]
nan表示不是一个数字not a number
inf表示无穷大infinity
#形状不完全相同计算，要么列形状相同，要么行形状相同
t1 = np.arange(12).reshape(3,4)
print(t1)
>>>
[[ 0 1 2 3]
[ 4 5 6 7]
[ 8 9 10 11]]
#含有4个元素的一维数组
t2 = np.array([1,2,3,4]) #t2 = np.array([1,2,3,4]).reshape(1,4) 和一维数组效果一样
print(t2)
>>>
[1 2 3 4]
#t1每行和t2的列数量相同，列对应元素 做减运算
print(t1-t2)
>>>
[[-1 -1 -1 -1]
[ 3 3 3 3]
[ 7 7 7 7]]
#3行1列
t3 = np.array([1,2,3]).reshape(3,1)
print(t3)
>>>
[[1]
[2]
[3]]
#t1每列和t3行数量相同， 行对应元素 做减运算
print(t1-t3)
>>>
[[-1 0 1 2]
[ 2 3 4 5]
[ 5 6 7 8]]
#形状完全不同计算报错，行列形状都不同
t4 = np.array([1,2,3,4,5])
print(t1-t4)
>>>
ValueError: operands could not be broadcast together with shapes (3,4) (5,)
