A database that stores data by column, strong at aggregate processing.
An approach that physically stores data by column rather than by row, letting aggregate operations over a particular column run quickly. Because values within the same column tend to share a type and pattern, they compress well, and an aggregate query can skip columns it doesn't need entirely -- cutting disk I/O itself, which is the core of its performance edge over a row-oriented database.
© 2026 ITBGM