Location-based query processing has been an active research area for years. With the development of the location-aware applications and services, users’ queries and needs are becoming more complex. A user’s query is a combination of multiple location-based and relational operations, e.g., location-based selection, location-based joins, relational selection, and relational joins. Current location-based query systems provide only one location-based operator, e.g., range or k-Nearest-Neighbor operators. Our focus is to improve location-based query systems to correctly interpret and efficiently evaluate combinations of location-based and relational operations. PostgreSQL is a widely adopted open-source and extensible relational database management system. The objective of this thesis is to extend PostgreSQL to process combination of location-based operations intermixed with relational operations to handle complex user queries. The focus will be on studying the optimization techniques proposed in Spatial queries with k-nearest-neighbor and relational predicates for the kNN operators with relational predicates. In addition, support those optimized techniques in the context of PostgreSQL. Our conducted experiments show the enhancements of the execution performance of our implementation with three order of magnitude compared to the original kNN operators inside PostgreSQL.
This is a development version.