Ανάπτυξη μεθόδων χωρο-κειμενικής ευρετηρίασης σε μη-σχεσιακές βάσεις δεδομένων
Development of methods for spatio-textual indexing in NoSQL stores
View/
Keywords
Spatio-textual queries ; Spatio-textual indexing ; NoSQL stores ; MongoDB ; HBase ; Hilbert curve ; Boolean range queries ; Space-filling curveAbstract
Nowadays, many IT companies provide GPS services and products to their customers. The most distinctive example is Google via the Google maps app. The app’s fundamental service is to provide its users with directions from one point to another. The companies Uber and Beat also have apps to provide their customers transport services from one start point to any destination. One simple query that the Google maps app supports is the search of nearest cafes from a specific location. Uber and Beat give their customers the option to select the driver’s sex before transport. The conclusion is that all apps mentioned above can process both spatial and textual operating queries. In addition to spatial and textual data processing, they can process temporal data, but this case will not be examined in this study.
All apps in the previous paragraph store and process their data using Data Bases with the following specific features: high performance, availability and scalability. A Relational Database Management System (RDBMS) cannot cover all these needs, except for NoSQL Stores. However, some NoSQL Stores do not support direct spatial or spatiotextual indexing, even though they do have some techniques to support this issue.
This study will present spatio-textual techniques implemented on NoDA API for MongoDB and Hbase Stores. The NoDA API is an abstract layer between an app and NoSQL Stores, providing one query language and supporting spatial, spatio-temporal and spatio-textual indexing.