A. The CREATE CORTEX SEARCH SERVICE command requires that CHANGE_TRACKING = TRUE be enabled on the source table, especially if the role creating the service is not the table owner. This ensures that the service can track updates to the base data.

B. Snowflake recommends using a dedicated virtual warehouse of any size, including X-Large or 2X-Large, for each Cortex Search Service to ensure the fastest possible materialization of search indexes during creation and refresh.

C. Columns specified in the ATTRIBUTES field during service creation are only used for filtering search results and do not need to be present in the source query.

D. Cortex Search is designed to get users up and running quickly with a hybrid (vector and keyword) search engine on text data, handling embedding, infrastructure maintenance, and search quality parameter tuning automatically.

E. Cortex Search Services currently support replication and cloning, allowing for easy disaster recovery and geographical distribution of the search index.