So, here’s an idea: While CSV upload of data is great and everything, CSV is fast becoming outmoded vis-a-vis JSON representations of the same data.
Therefore, should we not be able to upload data and create database things from a pure JSON file? This would be extremely handy when building systems. Also, it’s just much simpler than mapping fields in a CSV. JSON is self-describing, so theoretically you wouldn’t even have to have a mapping step… Just, “Hey, Bubble, here are some things. Make these things.”
Specific use case: I’m currently creating some system-type Country objects based on https://github.com/mledoze/countries. While this dist includes a CSV that describes Countries, it is actually inferior to the JSON representation of same. Case in point: The names of countries are provided in various languages. However, in the CSV, these turn into a list of translations without indicators for what the target language is. In the JSON, this is made very specific.
So this is a request for an Upload type feature that accepts JSON.
There are (sort of) workarounds for this, like feed that data into the API connector, but you’d still have the issue of the API Connector’s wonky handling of lists of things and sublists of things.
It seems completely weird to take a thing that is JSON, turn it into CSV, and upload it to Bubble where essentially it will turn into JSON. Save us some steps, peeps! 
The SUPER COOL version of this wouldn’t even require a file. You’d just point Bubble at a URL for the JSON and bam, things created. (Obviously, in some cases, you’d have to provide a bit of guidance like if you want to “ignore” fields/keys, but you get what I’m saying.)
