What Spark/Delta Lake choose ORC vs Parquet file format? I learnt ORC is much faster when querying, It is much compression efficient than parquet and has most the feature which parquet has on top of it? Why not choose ORC? Am I missing something? Please help