* No special value source type. Otherwise a new aggregation is needed.

* A field mapper that includes duplicate values when aggregating on it. This is be done by using two doc values fields. One for the values and one for the counts.
* Wrapping sorted set docvalues instance to include duplicates in the count returned by docValueCount(...).
* Report the counter field mapper, otherwise merging may fail. No field can be unmapped.
* Currently doesn't work when filter by filter terms aggregations is enabled.