I'm trying to understand what takes place when a JSON1 extension function is used to update a property in a JSON string stored in a data table, compared to simply overwriting the row. The reason I am asking is that I'm moving a web extension's storage from the indexedDB API to a C utility using SQLite3 and the browser's native-messaging API. Because of the manner in which the garbage collector did not release RAM until a certain threshold had been consumed, it was more efficient to hold a certain amount of data in RAM and use it to overwrite the existing database records through a `put` operation only, as opposed to a `get/put` combination, in which the record was first retrieved, then de-serialized, updated, serialized, and put back in the database. Would something similar be true in SQLite using the JSON functions for an update, such that, to update some properties of the string, the record would be selected, parsed, updated, and written back; and just replacing the full entry would be less work? If true, then there appears to be two options: 1) pass the complete new JSON string from the browser to the utility and perform a less-intensive SQLite operation, or 2) pass only the data that needs updated and perform a more-intensive SQLite operation. Is this an accurate view of the SQLite side of the process? Thank you.