Support multi column support inside count(distinct c1,c2..) which is not working in Spark SQL.
[Github] Pull Request #3511 (ravipesala)