Best practices for statistics gathering in a small data warehouse
Are there any articles available that discuss best practices for gathering statistics for a data warehouse?
We're currently running 10.2.0.5 (11g next year) and our database team as automated statistics gathering with default values, which in my understanding is considering a table 'stale' if the number of rows inserted, updated or deleted is greater than 10% of the table. This seems inadequate for some of our tables. When we find reports running poorly we observe tablescans on medium or large tables. Running dbms_stats.gather_table_stats('owner','tablename') against the table usually results in the query plan changing to use an index and the report running in seconds vs minutes.