A schema evolves frequently and contains heterogeneous JSON values. Which current Databricks type should be evaluated?

1. A fixed BOOLEAN for all fields
2. VARIANT
3. A partition column for each possible JSON character
4. A single INT for the entire document