mutate() adds columns worked out from the ones you have, row by row: orders |> mutate(revenue = units * unit_price). You can create several in one call, and a later one can use an earlier one. Reusing an existing name replaces that column.
if_else(condition, yes, no) picks one of two values for each row: if_else(minutes > 30, "long", "short"). Both answers must be the same type.
Inside mutate(), a function such as sum() or mean() sees the whole column, so mutate(total = sum(units)) puts the grand total on every row. That is useful for shares, and a common accident when you meant a per-row value.
Missing values spread: a row with a missing discount gets a missing net price. Decide what a missing value means before you calculate with it.
Two new columns; the second uses the first.