关于numpy的简单操作
资源内容介绍
关于numpy的简单操作 '''# 创建数组arr = np.array([1,2,3,4,5,6,7,8,9])print(arr)arr1 = np.empty([2,3])print(arr1)arr2 = np.zeros([3,4])print(arr2)arr3 = np.ones([4,5])print(arr3)arr4 = np.zeros_like(arr1)print(arr4)arr5 = np.ones_like(arr1)print(arr5)arr6 = np.asarray([(1,2,3),(4,5,6)])print(arr6)arr7 = np.fromiter(range(10), dtype=int, count=5)print(arr7)arr8 = np.arange(10,20,2)print(arr8)arr9 = np.linspace(10,20,3)print(arr9)'''import numpy as np'''arr = np.array([0, 1, 22, 333, 4444, 555, 66, 7, 88, 999])# 切片和索引print(arr[slice(1, 7, 2)])print(arr[1:7:2])arr1 = arr.reshape(5, 2)print(arr1)print(arr1[0:2, 1:3])#布尔索引def fn(a): return a > 3print(arr1[fn(arr1)])print(arr1[~fn(arr1)])#花式索引arr2 = np.array([1, 3, 5, 9])print(arr[arr2])#二维数组#读取指定的行print(arr1[[1,4,2]])#倒序引入print(arr1[[-1,-4,-2]])#传入多个索引数组（要使用 np.ix_）,前面为第几行，后面为该行的第几个print(arr1[np.ix_([1,4,2],[1,0])])''''''# 广播：允许NumPy用不同大小的数组进行算术运算。# 维数相同，各维度长度相同a = np.array([1, 2, 3])b = np.array([4, 5, 6])c = a * bprint(c)# 形状不同时b = np.array([[4, 5, 6], [4, 3, 2]])c = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])print(c)# 数组的长度要相同,且至少有一个维度是一维d = a * cprint(d)# e = b*c# print(e)# ValueError: operands could not be broadcast together with shapes (2,3) (3,3)'''# 修改数组形状arr = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9])# reshape中数字相乘为arr的个数arr1 = np.reshape(arr, [3, 3])print(arr1)# flat为迭代器，可通过循环遍历数组for i in arr1.flat: print(i)# 返回拷贝数组，不修改原来的数组arr2 = arr.flatten()arr2 = arr2.reshape(3, 3)print(arr)print(arr2)#平铺数组,不修改原来的数组arr3 = arr2.ravel()print(arr3)print(arr2)#翻转数组,不修改原来的数组arr4 = np.transpose(arr2)print(arr4)print(arr2)#转置数组,不修改原来的数组arr5 = arr2.Tprint(arr5)print(arr2)#连接数组#numpy.concatenate 函数用于沿指定轴连接！！相同形状！！的两个或多个数组arr6 = np.concatenate((arr1, arr5), axis=1)print(arr6)#分割数组,indices_or_sections必须能划分为等量的份数arr7 = np.split(arr6,indices_or_sections=3,axis=0)print(arr7)#hsplit指定要返回的相同形状的数组数量来拆分原数组。只能水平分割arr8 = np.hsplit(arr6,2)print(arr8)#vsplit指定要返回的相同形状的数组数量来拆分原数组。只能垂直分割arr9 = np.vsplit(arr6,3)