The Research Of Aggregate Algorithm Over Data Streams

With the fast development of computer application, nowadays data stream processing is becoming the new hot field of database research. Aggregate query is a common query type in data stream query processing.In this paper, we focus on the aggregation query over the continuous data streams. This paper presents aggregate algorithms, which are based on the aggregate algorithms, using the idea of separating longitudinally into bucket. In these algorithms, a window is separated into buckets by using the greatest common denominator between the window and the hop. It implements the diversification of windows and meets the majority of requirements of queries. It mainly studies simple aggregate algorithms and group aggregate algorithms of various windows.In the aggregate algorithms, a window separates into buckets by using the greatest common denominator between the window and the hop, so computing the value of aggregate and maintenance of windows are according to the unit of the bucket. In the bucket, it merely saves the value of aggregate, but not tuples of the data stream. It reduces memory cost of query operators effectively. The simple aggregate algorithm of windows mentioned in this paper will increase query efficiency with the method of using current aggregate answer to compute the next aggregate answer; The group aggregate algorithm uses the method of hash. The tuple is saved according to group attributes. Theory analyses and experiment results show that the algorithms proposed in this paper have the good cost of time and space.