What happens to the underlying table data when a CLUSTER BY clause is added to a Snowflake table?

Question

Options

• AData is hashed by the cluster key to facilitate fast searches for common data values
• BLarger micro-partitions are created for common data values to reduce the number of partitions that
• CSmaller micro-partitions are created for common data values to allow for more parallelism
• DData may be colocated by the cluster key within the micro-partitions to improve pruning

• A
  2% (1)
• B
  3% (2)
• D
  95% (56)

Explanation

Adding a CLUSTER BY clause defines a clustering key but does not immediately reorganize data. Snowflake's Automatic Clustering service then progressively reorganizes micro-partitions over time so that rows with similar clustering key values are stored together (colocated) within the same or adjacent micro-partitions. This colocation significantly improves partition pruning - queries filtering on the cluster key can skip large numbers of irrelevant partitions. Data is not hashed (that would be a hash distribution model). Micro-partition sizes remain consistent; the goal is value colocation, not size adjustment.

No community discussion yet for this question.