Google's crash course practice first_steps_with_tensor_flow step 5 machine learning code:

_=linear_regressor. "train" (
Input_fn=lambda: my_input_fn (my_feature, the targets),
Steps=100
)

Here the training function return value assigned to a variable _, but behind the code does not use the _ variable, then why not just write here:

Linear_regressor. "train" (
Input_fn=lambda: my_input_fn (my_feature, the targets),
Steps=100
)

So what is the purpose of "redundant"?