Repetitive queries refer to those queries that are likely to be executed repeatedly in the future. Examples of repetitive queries include those that are used to generate periodic reports, perform routine maintenance, summarize data for analysis, etc. They can constitute a large part of daily activities of the database system and deserve more optimization efforts. In this paper, we propose to collect information about joins of a repetitive query, called the trace, during execution. We intend to use this information to re-estimate selectivities of joins in all possible execution orders. We discuss the information needed to be kept for the joins and design an operator, called the extended full outer join, to gather such information. We show the sufficiency of the traces in computing the exact selectivities of joins in all plans of the query. With the exact selectivities of joins available, the query optimizer can utilize them to find truly the best join order for the query in its search space, guaranteeing "optimal" execution of the query in the future.
展开▼
