Parallel execution of aggregates

In aggregate-only queries, the database server can break the computation of the aggregate into several pieces and compute each piece in parallel. The database server then uses the COMBINE function to combine the partial results from all pieces in a single result value. The database server uses the optimizer to decide when and how to parallelize an aggregate. This action is transparent to the user.

In queries that are not exclusively aggregate, the database server can still compute multiple aggregate results in parallel. In such cases, the database server computes each aggregate result sequentially (without using the COMBINE function).

For more information about parallelization and optimization, refer to the HCL OneDB™ Performance Guide.