我不明白为什么 np.argwhere 没有为我的第一行返回一些东西

Question

I have an array, and I want to find where are the max value for each row. On the same row I can have the two values that are the maximum, so I decided to use np.argwhere.

np.argwhere(a == np.amax(a,axis=1))

but for my first row, it output nothing and I don't understand why.

Answer 1

np.amax returns a row array, while you want a column one. Hence, you want to reshape your array before performing the computation:

np.argwhere(a == np.amax(a, axis=1).reshape((-1,1)))