medium10 pts · first answer countsparquetfile-formatsperformance

How Parquet skips data

A query filters WHERE order_date = '2026-10-01' on a Parquet file. How can the reader avoid reading most of the file?

A.Parquet files are always read in fullB.Parquet stores a hash index of every rowC.Footer metadata stores min and max values per column chunk in each row group, so non-matching row groups are skippedD.The file name contains every date in it

← Eight consumers, six partitions Undoing a bad load →