在实际操作中,我们需要需要养成习惯,任何浪费空间或者时间的操作都需要减少,因为我们操作的是大数据,可能造成灾难性的放大,使用numpy的思想,尽量减少不必要的拷贝,我们操作的对象是大数据。
import numpy as np
x = np.array([1,2,3])
print(x)
x1 = np.array([[1,2,3],[4,5,6],[7,8,9]])
print(x1)
print(x1[:,0])
x2 = np.random.random((3,3,3))
print(x2)
print(x2[:,:,0])
x2[:,:,0] =0
print(x2[:,:,0])
print(x2)
print(x1)
print(np.average(x1,axis=0))
print(np.average(x1,axis=1))
import time
x = np.random.random((1000,1000))
ans =np.zeros((1000,1000))
t = time.time()
for i,row in enumerate(x):
for j,elem in enumerate(row):
ans[i][j] = elem +1
print(time.time()-t)
t = time.time()
ans = x+1
print(time.time()-t)
t = time.time()
np.add(x,1,ans)
print(time.time()-t)
x = np.array([1,2,3])
y = np.array([[1,1,1],[2,2,2],[3,3,3]])
z = y-x
print(z)
print(end ='\n')
x =y = np.array([1,2,3])
y = np.tile(y,[3,1]).T
z = y-x
print(z)
print(end ='\n')
x =y = np.array([1,2,3])
t = y[:,None]
print(t)
z = t -x
print(z)
print(end ='\n')