Data Expiration and Aggregate Queries

We investigate the space requirements for summaries needed for maintaining exact answers to aggregate queries over histories of relational databases. We show that, in general, a super-logarithmic lower bound (in the length of the history) on space needed to maintain a summary of the history in order to be able to maintain answers to counting queries. We also develop a natural restriction on the use of aggregation that allows for a logarithmic upper bound and in turn maintaining the summary of the history using counters.

• Type: Conference
• Year: 2009
• Where: AMW
• Authors: David Toman
• Keywords: AMW 2009, AMW 2010, Exact Answers, Logarithmic Upper Bound, Super-logarithmic Lower Bound