DuckdbAug 10, 2025 TIL: DuckDB's COPY reads from S3 URLs directlyPoint DuckDB at an s3:// URL and it will read Parquet files without downloading them first.