Apache ORC is a self-describing type-aware columnar file format designed for
Hadoop workloads. It is optimized for large streaming reads, but with integrated
support for finding required rows quickly.