Which type of tables typically benefits from creating a cluster key?

• A. Small tables (around 1GB)
• B. Medium tables (around 10GB)
• C. Large tables (around 100GB)
• D. Very large tables (multi-terabytes of data)

Answer: (D)

Clustering keys are most beneficial when you have very large tables because they enable Snowflake to prune micro-partitions based on the clustering column values, reducing the amount of data that needs to be scanned for filtered queries. As a table grows into multi-terabyte scale, it consists of many micro-partitions; without a clustering key, a query with a filter may have to touch a large portion of those partitions. A clustering key aligns the data so that the relevant partitions can be skipped, which can dramatically lower compute costs and improve response times.

For small tables, the number of partitions is already limited, so the potential pruning benefit is minimal and the overhead of maintaining a clustering key isn’t worth it. Medium-sized tables might see some gains if their queries consistently filter on the clustering columns, but the most pronounced and reliable benefits occur with very large datasets.