Since its release in 2013 as a columnar storage for Hadoop, Parquet has become almost ubiquitous as a file interchange format that offers efficient storage and retrieval. This adoption has led to it becoming the foundation for more recent data lake formats, e.g., Apache Iceberg. In this blog series, we explore how ClickHouse can be used to read and write this format before diving into the Parquet in more detail. For more experienced Parquet users, we also discuss some optimizations that users can make when writing Parquet files using ClickHouse to maximize compression, as well as some recent developments to optimize read performance using parallelization.
For our examples, we utilize the UK house price dataset. This contains data about prices paid for real estate property in England and Wales from 1995 to the time of writing. We distribute this in Parquet format in the public s3 bucket s3://datasets-documentation/uk-house-prices/parquet/.