Database system providing optimization of group by operator over a union all

摘要: A database system providing improved methods for optimization and execution of queries aggregating information from multiple tables is described. The method breaks serial operations into a group child that can be individually optimized processed in parallel. query including grouping operator over union node split by pushing the operation down each input arm node. Specifically, operators are inserted qualifying nodes, enabling separate physical to summarized Each concurrently groups data generates summary results parallel, taking advantage capabilities available parallel processing hardware. aggregate original master modified accept correctly synthesize sub-aggregated generated operators. also enables individual on table, use advantageous index-only plans.