DEVELOPMENT OF SQL INTERFACE FOR SPATIAL DATA PROCESSING IN XML DATABASES

• Main Author: Bintang Pananjung, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/76206
• Institution: Institut Teknologi Bandung

With the advancement of technology, the utilization of spatial data has become increasingly prominent in applications facilitated by the Global Positioning System (GPS). Furthermore, spatial data is often stored in document data structures such as GeoJSON, GML, and KML. However, Database Management Systems (DBMS) that support document data structures, such as document databases and XML databases have limited spatial operations. Pradipta (2020) proposed the use of the PostGIS extension for MongoDB to complement spatial operations, subsequently developed further by Adzkiya (2021) to provide abstractions within document- based databases with several optimizations. Nevertheless, Adzkiya's solution (2021) has not yet transformed join and aggregation queries, nor has it handled the difference in data structure between document structure and SQL structure, such as metadata and nested collections. Furthermore, differences in the capabilities of spatial data processing also need to be addressed within the XML DBMS for each respective version. In this final project, an SQL interface as Adzkiya's solution (2021) is adapted for implementation by integrating the XML database into the extension of the abstraction. Handling of joins and aggregations is accomplished by modifying the Abstract Syntax Tree, which is processed for XML database operations. Additionally, abstractions are developed for each spatial data format (KML and GML) as well as for different versions of the XML DBMS. On the other hand, additional information within the XML document structure, such as attributes and nested nodes/collections, needs to be addressed. This is achieved through the transformation of XML queries using mechanisms like attribute flattening and Cartesian product between nested collection and document in the main collection. Evaluation is conducted by examining the correctness of results and the performance of query executions implemented in this final project. The correctness of results is assessed based on the comparison of data results and data expected, while performance evaluation involves testing queries with large amounts of data. The evaluation results conclude that the correctness of the results aligns with expectations. Furthermore, the execution with large data volumes has effectively enhanced the efficiency of rebuilding queries by minimizing the amount of data extracted from the XML database.