For anyone else with more time to roll up their sleeves, https://github.com/ClickHouse/ClickHouse/blob/b0ddc4fb30f1a0... would be the place to split the header row, and perhaps use some heuristics on the first data row to identify datatypes!
For anyone else with more time to roll up their sleeves, https://github.com/ClickHouse/ClickHouse/blob/b0ddc4fb30f1a0... would be the place to split the header row, and perhaps use some heuristics on the first data row to identify datatypes!