Aggregators not generic


I am new to Druid. I am trying to convert a HQL (Hive query) into a Druid Query. I notice that , in case of aggregators, there are different implementations for sum, min and max for different datatypes…like longsum, doublesum, doublemin aggregator factories etc. Aren’t there generic sum , min and max aggregators available ? I don’t have the datatypes of the columns in my input and need to make explicit calls to the database to find the datatype of all aggregated columns everytime. Is there a way to avoid this ?



The type is the output type rather than the input type. Whatever is stored there will be converted to the output type before aggregation. So, hopefully that makes it simpler.