Επεξεργασία χωροχρονικών ερωτημάτων στην MongoDB
View/ Open
Keywords
MongoDB ; Spatio-temporal data ; Βάσεις δεδομένων ; Αλγόριθμοι ; Αρχιτεκτονική συστήματοςAbstract
The main subject of this thesis was query processing of spatiotemporal data in the NoSQL database MongoDB. The definition for spatiotemporal data is as follows: data objects that combine coordinates (latitude, longitude) with a timestamp, which represent when the entity under examination was present at the given time in the given geographical point. In addition to this data, there exists a unique identification (id) for each entity. This data was stored in the MongoDB database, with each document containing the id, latitude, longitude and the related timestamp. The queries answered were the Circle Range query and the Box Range query, regarding finding data in a circle or a rectangle for a specific period of time [t_min, t_max] respectively. Also, the k-Nearest Neighbours query and k-distinct-id-Nearest Neighbours query, regarding finding the closest coordinates, and the closest coordinates that belong to distinct entities near to a specific point (x,y) in a given period [t_min, t_max] respectively. In the experimental part of this thesis, different metrics were used: the time needed to execute the query, the time of algorithm execution in total and the number of the documents accessed for the query to be answered. Regarding the use of indexes, all the queries were processed: without any index, using index in the field coordinates, using index in the field time, and using compound index which combines the fields coordinates and time.