Given that you only have two columns, something like

df.group_by("Letter").all()

would also give you the desired result. If you have more columns, all() would turn everything except for the group_by column into series of lists.