Statistics for columnstore indexes
Because the characteristics of a data warehouse workload are typically very different from those of a transaction workload, the methodology you follow for the maintenance of statistics will depend on the characteristics of your data. As you develop your plan, consider the following guidelines:
- For columns that contain static values, such as certain dimension tables, reduce the frequency of updates to the statistics.
- An ascending key column for which new values are added frequently, such as a transaction date or an order number, likely requires more frequent updates to the statistics. Consider updating the related statistics more often.
- Consider using asynchronous statistics for workloads, such as data warehouses, that frequently execute the same query or similar queries. Query response times could be more predictable because the Query Optimizer can execute queries without waiting for up-to-date statistics.
- Consider using synchronous...