SEARCH RESULT

In modern database management systems, optimizing query performance is a critical requirement, especially for large scale non-spatial data. Traditional indexing techniques use data structures such as B-Trees to optimize queries on non-spatial data. However, since B-Trees can organize data only on a single column, they cannot efficiently process complex queries involving multiple attributes. To handle such queries, database designers use composite indexes on attributes (columns) that are frequently queried together. However, composite indexes also have certain limitations and inefficiencies. In composite indexes, each column is sorted based on the column before it; therefore, if the previous column is not included in the query, the index becomes inactive. Consequently, to utilize a composite index, unnecessary columns must also be included in the query, making it inefficient. In this study, we propose a novel approach using PostgreSQL's cube extension to transform multiple non-spatial attributes into multidimensional data. Each non-spatial attribute represents a dimension in the space. This transformation enables the use of spatial indexing methods such as R-Tree, significantly improving query performance. At the same time, it eliminates the limitations of composite indexing. Particularly in large scale databases, this approach provides substantial performance improvements in filtering and range queries, offering significant advantages over traditional indexing techniques. By indexing and querying non-spatial data as multidimensional objects, this method achieves a query performance that increases linearly compared to traditional methods as the query range grows in range queries. This study provides a practical and effective solution for databases handling complex multi-attribute queries, bridging the gap between spatial and non-spatial indexing techniques in terms of query performance and flexibility

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Enes Özdeniz Hazim İşcan