Currently not loading large datasets into memory using the option use_memory_table: false doesn’t work for CSV. It works fine for parquet as mentioned in the documentation: https://roapi.github.io/docs/config/dataset-formats/parquet.html#large-datasets