Frontend <> Backend communication about desktop HOT 6 CLOSED

I don't think it is a good idea to manually build the query AST in JS land. It couples the Frontend and Backend internals too much. Sending a SQL statement feels cleaner.

Yes, building the AST in JS only made sense as long as the AST could immediately be executed there too.

from desktop.

I vote for Option 1 too.

I additionally vote for using JSON as the output as it's easier to use in the frontend than parsing some CSV output.

We can live without progress reports, and if we need it later we can get hacky then.

from desktop.

Do we drop the JS CSV parser? If yes, do we use the backend to figure out the shape when we load a file?
If not, we need to use 2 different CSV libraries where each may have their own tiny differences.

from desktop.

Do we drop the JS CSV parser? If yes, do we use the backend to figure out the shape when we load a file?
If not, we need to use 2 different CSV libraries where each may have their own tiny differences.

Good point, @edongashi.

We either need another parser for the GUI or need to extend the Reference with an endpoint that returns a schema...
In either case, as long as we want to support CSV, it seems the CLI interface must be extended to support providing a schema as part of the input too!?

from desktop.

I say we do the CSV parsing only in the backend/reference tool.
To load the initial raw data (including the schema) the frontend could issue a standard SELECT * FROM 'file_name' query.

from desktop.

This seems settled (at least for now).

from desktop.