numpy中keepdims用法

keepdims=True(默认False)

保持矩阵维数不变(被减少的那个维度会以维度1保留在结果中)。

 

import numpy as np

>>> a = np.random.randint(10,size=(3,5))
>>> a
array([[6, 4, 5, 4, 4],
       [9, 6, 9, 2, 3],
       [9, 2, 9, 1, 1]])

# 按行相加
>>> np.sum(a,axis=1)
array([23, 29, 22])
>>> np.sum(a,axis=1,keepdims=True)
array([[23],
       [29],
       [22]])

#按列相加
>>> np.sum(a,axis=0)
array([24, 12, 23,  7,  8])
>>> np.sum(a,axis=0,keepdims=True)
array([[24, 12, 23,  7,  8]])