S3: Stats gathering for insert queries can be expensive for partitioned dataset

StatsTask's aggregateStats populates stats details for all partitions by checking the file sizes which turns out to be expensive when larger number of partitions are inserted.