This paper introduces a new interface for online aggregation in database systems, which allows users to observe and control the progress of aggregation queries in real-time. Traditional batch processing is often frustrating for users due to its lack of feedback and long response times. The proposed online aggregation interface provides users with control over the execution, such as the rate at which different aggregates are computed and the trade-off between accuracy and speed. The paper outlines usability and performance requirements for such a system and presents techniques to extend a relational database system to support online aggregation. These techniques include methods for returning output in random order, computing running confidence intervals, and modifying access methods to ensure random access of data. The authors also discuss the importance of statistical estimation techniques and how they can be incorporated into the system to help users assess the precision of running aggregates. An initial implementation of online aggregation in POSTGRES is described, along with performance results that demonstrate the effectiveness of the proposed techniques. The paper concludes by highlighting the benefits of the online aggregation interface and suggesting future work in extending the SQL API to support more interactive query control.This paper introduces a new interface for online aggregation in database systems, which allows users to observe and control the progress of aggregation queries in real-time. Traditional batch processing is often frustrating for users due to its lack of feedback and long response times. The proposed online aggregation interface provides users with control over the execution, such as the rate at which different aggregates are computed and the trade-off between accuracy and speed. The paper outlines usability and performance requirements for such a system and presents techniques to extend a relational database system to support online aggregation. These techniques include methods for returning output in random order, computing running confidence intervals, and modifying access methods to ensure random access of data. The authors also discuss the importance of statistical estimation techniques and how they can be incorporated into the system to help users assess the precision of running aggregates. An initial implementation of online aggregation in POSTGRES is described, along with performance results that demonstrate the effectiveness of the proposed techniques. The paper concludes by highlighting the benefits of the online aggregation interface and suggesting future work in extending the SQL API to support more interactive query control.