Why might we create a 'train_df' /'test_df'?

a) To confuse the model during training.
b) To store intermediate results.
c) To separate training and testing data.
d) It is a standard naming convention.