1997 | Joseph M. Hellerstein, Peter J. Haas, Helen J. Wang
Online aggregation is a new interface that allows users to observe and control the progress of their aggregation queries in real time. Traditional database systems perform aggregation in batch mode, which is inefficient and unresponsive to user needs. This paper proposes an online aggregation interface that provides running estimates, confidence intervals, and control over the execution of aggregation queries. The interface allows users to monitor the progress of their queries and adjust the processing based on their needs. The system supports a variety of aggregation functions and provides statistical estimation techniques to help users assess the accuracy of running aggregates. The paper also discusses the performance and usability goals of online aggregation, including the need for efficient data access methods, fair and non-blocking group processing, and non-blocking join algorithms. The implementation of online aggregation in POSTGRES is described, along with performance results that demonstrate the effectiveness of the system. The paper also presents formulas for computing confidence intervals and discusses the benefits of online aggregation for interactive data exploration and analysis. The results show that online aggregation provides a more efficient and user-friendly approach to aggregation than traditional batch processing.Online aggregation is a new interface that allows users to observe and control the progress of their aggregation queries in real time. Traditional database systems perform aggregation in batch mode, which is inefficient and unresponsive to user needs. This paper proposes an online aggregation interface that provides running estimates, confidence intervals, and control over the execution of aggregation queries. The interface allows users to monitor the progress of their queries and adjust the processing based on their needs. The system supports a variety of aggregation functions and provides statistical estimation techniques to help users assess the accuracy of running aggregates. The paper also discusses the performance and usability goals of online aggregation, including the need for efficient data access methods, fair and non-blocking group processing, and non-blocking join algorithms. The implementation of online aggregation in POSTGRES is described, along with performance results that demonstrate the effectiveness of the system. The paper also presents formulas for computing confidence intervals and discusses the benefits of online aggregation for interactive data exploration and analysis. The results show that online aggregation provides a more efficient and user-friendly approach to aggregation than traditional batch processing.