How can I remove the 0 in the 2-dimensional tensor?

Greetings, I want to extract the nonzero elements in the 2-dimensional tensor.

example
# tensor([[    0,  4535,     0,  ...,  9446,  5344, 0],
#         [    0,  1037,     0,  ...,  2391, 33243, 40207]])

result
# tensor([[4535,   ...,  9446,  5344],
#         [1037,  ...,  2391, 33243]])

I thought it was a simple problem, but I countered the error because of the mismatch of shape.
I tried as below. Does anyone have a good idea?

result[:, torch.cat((result[0].nonzero()[:, 0], result[1].nonzero()[:, 0])).unique()]

#tensor([[ 4535, 34602, 31080,  2097, 30280, 30764,     0, 30471,  5728, 32816, #Unexpected 0 
#              2012,     0,  6726, 30444,     0, 34984, 32871, 32039,     0, 32528, ...], #Unexpected 0 
#            [ 1037, 35914,  9275,  7511,  9624,  9925,  7217,  9039,  1386,   362,
#              2022,  2719,     0,  6996, 31594, 34346, 33777, 34213, 37084, 32728,...]]) #Unexpected 0 

Thank you for reading this question.