Defining a clustering key directly on top of VARIANT columns is not supported; however, you can specify a VARIANT column in a clustering key if you provide an expression consisting of the path and the target type.

• A. False
• B. Only if you specify a path
• C. True
• D. Only for arrays

Answer: (C)

Clustering keys use scalar, typed values, not complex semi-structured data. A VARIANT column holds nested data, so you can’t define a clustering key directly on it. But you can include a VARIANT column in a clustering key if you extract a specific path inside that VARIANT and cast the result to a concrete type. This path plus the target type forms a deterministic, comparable value that the clustering mechanism can use. For example, you could define a clustering key using a path like variant_col:'customer.id'::VARCHAR or variant_col:customer.id::VARCHAR, depending on syntax, to pull out a string field from the JSON and treat it as the clustering key. This matches the idea that you must specify both the path and the target type.