The distinct count is the number of unique values in a column or expression.

Kyvos supports two types of distinct count: Approximate and Accurate. Using the Approximate count improves performance, whereas the Accurate count increases the size and processing time of the semantic model.

The approximate distinct counts are computationally less expensive as compared to accurate ones. There can be many business use cases that do not need accurate distinct counts, or it is acceptable to trade off accuracy over performance. Some businesses also want to keep the size of the semantic model optimal. In such cases, it is preferred to use Approximate distinct counts.