I am using Apache Spark and parquet-based external tables created via
CREATE TABLE IF NOT EXISTS {table_name} ({columns}) USING PARQUET {part_clause} LOCATION '{path}'