The correct answer is B because it accurately describes how clustering keys work in Snowflake. Clustering keys sort designated columns over time to optimize query performance by organizing data in micro-partitions, and this process is non-blocking - it doesn't interfere with DML operations. This is confirmed by the community discussion where all comments unanimously select B with upvotes, and references to Snowflake documentation support that automatic clustering is transparent and doesn't block DML statements. Option A is incorrect because clustering doesn't update micro-partitions in place with a full sort and doesn't significantly impact DML operations. Option C describes a different concept (likely indexing structures in other databases). Option D describes hash distribution for joins, which is not how clustering keys function in Snowflake.