QUERY HANDLING ON MOVING OBJECT DATABASES WITH DIFFERENT DATABASE SCHEMA AND REPRESENTATIONS OF OBJECT MOVEMENTS

• Main Author: Dicky Andika Putra, Mohammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/61869
• Institution: Institut Teknologi Bandung

A moving object database is a database that records the movement of an object over the time. There are various ways to build a moving object database. There are already database extensions capable of handling moving object databases, namely stpqueries which stores movement of an object in one attribute (attribute timestamped), and stpqueries-1nf which stores moving data in pairs of timestamp and value attributes (tuple timestamped). These two extensions have different representation models in storing moving data. In this final project, a system is built that can be used to retrieve information from a number of moving object databases with different database schemas and object movement representations using one SQL command. The system works by utilizing a metadata to map the global schema of the user's moving object database with a number of source database schemas. The commands that have been entered are translated so that they can be executed on each database source. This system is intended for retrieval of moving object data in the PostgreSQL database. The implementation is done by using node.js with the help of the node-sql-parser library to convert the input query form into an Abstract Syntax Tree (AST) and vice versa. The test includes several test cases of SQL queries to access ordinary data, as well as queries to access moving object data, whether the data exists in only one database, or data that is spread over several database sources. The test results show the expected output, with certain limitations due to the tools used.