![]() ![]() ![]() |
![]() |
|
|
![]() ![]() ![]() ![]() ![]() |
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Return to Tutorials There has been recent interest in techniques for online query processing, which produce incrementally refining results for longrunning queries. In online aggregation queries, incremental results are provided via estimations of the final results, along with statistical confidence intervals for the estimations. In online enumeration queries, incremental results can be provided in the form of individual rows or fields of an output relation. In both cases, these results are perceived by the user over time, and hence their improvement in quality should be maximized over time - a goal that is quite different from traditional metrics of system performance. As a corrolary to this interactive feedback, it is natural to provide users of online query processing systems with control over a system while it runs, allowing them to change their performance and query specifications based on the feedback they receive from the system. This research agenda brings together a number of threads in database research, including query processing architectures and algorithms, sampling and statistical estimation, and user interfaces for data visualization and transformation. ![]() DiSC'02 © 2003 Association for Computing Machinery |