Add support for multi-level pivot tables. It's now possible to pass a list to the `rows` parameter to get results where the rows are grouped by more than one column. E.g,
>>> pivot_table = pivot(ShirtSales, ['region', 'gender'], 'shipped', 'units')
To see the sum of units on each shipped date grouped by both region and gender.
Add the ability to change the row keys by passing a function that transforms a string.