But in essence what I meant is (1) A better API to access the underlying relations (2) surface this better on the UI.
For instance, when we upload a file using duckdb wasm iit creates a table, but if we create a table from a remote data source it creates a table and we are not tracking it (i.e. what is there on the left panel is currently is coupled with uploading files and not tables, whereas we should be tracking the tables)
Another question worth asking is do we store the results also as a temporary table, not sure how this would look like in terms of structure seems interesting nonetheless.