Which statement about clustering in Snowflake is true?

• A. Clustering speeds range searches and equality searches as long as they are on the clustering key.
• B. A table cannot be clustered on more than one key.
• C. Clustering automatically updates with every DML operation without maintenance.
• D. Clustering improves only write performance.

Answer: (A)

Clustering in Snowflake is about organizing data with a clustering key so the system can prune micro-partitions during query execution. When a query includes predicates on the clustering key, Snowflake can skip partitions that cannot contain matching rows, which reduces the amount of data scanned and speeds up the query, especially for range searches and exact-match filters on that key. That’s why the statement about clustering speeding range searches and equality searches on the clustering key is the best description of how clustering improves performance.

You can define a clustering key on a single column or on multiple columns, so clustering on more than one key is possible — that’s why the idea that you can’t cluster on more than one key isn’t accurate.

Clustering doesn’t automatically update with every DML operation in real time. There is an option for automatic clustering, but it runs asynchronously and may not reflect every single change instantly.

Finally, clustering mainly improves read performance by pruning data; it does not only boost write performance, so that aspect isn’t correct.