Ad-hoc Top-k Query Answering for Data Streams

A top-k query retrieves the k highest scoring tuples from a data set with respect to a scoring function defined on the attributes of a tuple. The efficient evaluation of top-k queries has been an active research topic and many different instantiations of the problem, in a variety of settings, have been studied. However, techniques developed for conventional, centralized or distributed databases are not directly applicable to highly dynamic environments and on-line applications, like data streams. Recently, techniques supporting top-k queries on data streams have been introduced. Such techniques are restrictive however, as they can only efficiently report top-k answers with respect to a pre-specified (as opposed to ad-hoc) set of queries. In this paper we introduce a novel geometric representation for the top-k query problem that allows us to raise this restriction. Utilizing notions of geometric arrangements, we design and analyze algorithms for incrementally maintaining a data se...

• Type: Conference
• Year: 2007
• Where: VLDB
• Authors: Gautam Das, Dimitrios Gunopulos, Nick Koudas, Nikos Sarkas
• Keywords: Data Streams, Database, Top-k Queries, Top-k Query, VLDB 2007