~~~Moto "the simplest solution is almost always the best" guided us in the sense of data storage, thus it is essential to not overload the database with deep decompositions to simplify the SQL-queries. For example, for the class I and class II methanol masers, the actual observational data (including non-detections) is stored in the tables met1_data and met2_data. The interferometric maser spot data for both class I and class II masers is stored in the table maser_spots to prevent overload of the main observation table. The interferometric spot table is connected to the observation tables with a simple integer identifier obs_id. Result of observations grouping (done with DBSCAN algorithm) is stored in the view all_objects and is recalculated with each database update. The grouping table is connected to the main observation table with grp column, identifying a source group in format of galactic coordinates "G55.371+0.185". External data (including fluxes) is stored locally in the following tables: iras, akari, bgps, atlasgal, ego and a few others. The table transitions stores the known methanol maser transitions, table papers stores the paper information (abstract, reference, data type, etc.), and table stat stores the statistics (number of detections, etc.).
Read our paper for more information:
Online Database of Class I Methanol Masers
D. A. Ladeyschikov, O. S. Bayandina, and A. M. Sobolev
2019, The Astronomical Journal, 158, 233