What is a key benefit of using columnar formats like Parquet or ORC in AWS Athena?

In Athena, queries are charged based on the amount of data scanned. Parquet and ORC are columnar storage formats that store data by column rather than by row. Because a query typically needs only a subset of columns, Athena can skip reading entire columns that are not referenced, dramatically reducing the amount of data read. Additionally, these formats compress data more efficiently than row‑based formats such as CSV or JSON, which further reduces the size of the data that must be scanned. The combination of selective column reads and better compression leads to lower query latency and lower cost per query. For example, if a table has 10 columns and a query only needs two of them, a columnar format can read only those two columns, whereas a row‑based format would read all 10 columns for every row. This is why columnar formats are the recommended choice for large analytical workloads in Athena. In practice, this means that a 100 GB CSV file might be stored as a 20 GB Parquet file, and a query that scans only a few columns could read just 2 GB instead of 100 GB. The cost savings scale with data size and query complexity, making columnar formats essential for cost‑effective analytics on Athena.