When a data tree is tightly coupled (like a complex sample of nested data with some arrays from a sensor) and the entire tree is treated like a single thing by writes, the JSON column just keeps things easier. Reads can be accelerated with indexes as demonstrated here.
No comments yet.