I am a firm believer that data, once collected, should never be thrown away unless there is a very good reason for doing so. Storage is cheap these days. My collection of well over 31 thousand images (as of April 2023) takes up 269 gigabytes, a few percent of one of today's commodity disks. Some of the most useful collections of astronomical data are useful precisely because they have saved their contents for future re-analysis, quite possibly in search of phenomena which had not been discovered at the time of data acquisition and/or using methods which had not been invented back then. The Harvard plate collection, which goes back well over a century, is perhaps the best known example. Even older databases (which are still used by professional astronomers) include far-eastern records of comets, novae and supernovae, and Babylonian star catalogues and eclipse records.
Accordingly, I have developed a database schema for the PostgreSQL relational database, together with software for populating it and extracting information from it. The schema is very general and powerful, in that it stores a great deal of information about any particular image, the equipment used to take it, objects located within it, and so on. Unfortunately, the schema and software is not yet ready for release, in large part because of its power and generality. It is sufficiently complex that a serious documentation effort must be made before it is comprehensible to others. Even I forget important details some of the time!
Another database, not as complex as the above but still woefully undocumented, stores information specific to the observation and photometry of variable stars.