Split a time-ordered dataset

Implement split_history(rows, train_count).

rows is already ordered oldest to newest. Return [training_rows, test_rows], taking the first train_count rows for training. train_count is an integer in [0, len(rows)]. Keep order and do not modify rows. Return two separate lists.

Before coding, predict one result by hand. After the tests pass, explain your approach and add a boundary example of your own.