AggregateFunction class

The AggregateFunction class performs the aggregation. It computes values on each database node where relevant data is stored and then combines the results from the nodes. You must implement the following methods:

• initAggregate() - Initializes the class, defines variables, and sets the starting value for the variables. This function must be idempotent.

• aggregate() - The main aggregation operation, executed on each node.

• combine() - If multiple invocations of aggregate() are needed, Vertica calls combine() to combine all the sub-aggregations into a final aggregation. Although this method might not be called, you must define it.

• terminate() - Terminates the function and returns the result as a column.

The AggregateFunction class also provides optional methods that you can implement to allocate and free resources: setup() and destroy(). You should use these methods to allocate and deallocate resources that you do not allocate through the UDAF API (see Allocating resources for UDxs for details).

API

Aggregate functions are supported for C++ only.

The AggregateFunction API provides the following methods for extension by subclasses:

virtual void setup(ServerInterface &srvInterface,
        const SizedColumnTypes &argTypes);

virtual void initAggregate(ServerInterface &srvInterface, IntermediateAggs &aggs)=0;

void aggregate(ServerInterface &srvInterface, BlockReader &arg_reader,
        IntermediateAggs &aggs);

virtual void combine(ServerInterface &srvInterface, IntermediateAggs &aggs_output,
        MultipleIntermediateAggs &aggs_other)=0;

virtual void terminate(ServerInterface &srvInterface, BlockWriter &res_writer,
        IntermediateAggs &aggs);

virtual void cancel(ServerInterface &srvInterface);

virtual void destroy(ServerInterface &srvInterface, const SizedColumnTypes &argTypes);