Understanding parquet a bit more

While experimenting with a Spark job to read the same dataset in both CSV and Parquet formats, I observed that queries on Parquet were significantly faster. I was aware that Parquet, being a columnar storage format, is designed to be performant, but ...