Allow specifying (via CsvSchema) a way to map "extra" columns into specific key (to use via any setter)

[cowtowncoder]

(note: sort of related to #95)

As an alternative to being able to ignore "extra" columns (values for columns that come after all defined columns), it would make sense to allow these to instead be mapped to sort of virtual properties.
But since it is tricky to try to generate specific naming convention (and, for specific names, one could just create schema anyway), it should probably be enough to just let a single name be specified, and this name exposed. Doing this would allow this extra information to flow into any-setter (@JsonAnySetter).