```json { "table_name": "classification_candidates", "primary_key": "id", "candidates": [ { "rank": 0, "partition_key": "table_classification_id", "partition_key_label": "primary_identifier", "partition_key_reasoning": "table_classification_id has RPUV=1.15 (each classification has ~1-3 candidate key recommendations), uniqueness=0.870, and Top 3 Coverage=1.52%. Semantically natural: groups all candidate key recommendations for the same table classification together, which is the primary access pattern for this metadata table.", "clustering_key": "rank", "clustering_key_label": null, "clustering_key_reasoning": "rank orders candidates by priority (0=primary recommendation, 1=first alternative, 2=second alternative). Monotonicity=0.131 and RPUV=197.00. Sorting by rank within each table_classification_id partition presents recommendations in their intended order.", "world_line": null, "session_column": null, "confidence": "HIGH", "notes": "This is a dimension table storing key selection recommendations for classified tables. Each row represents one candidate (partition_key, clustering_key) combination for a table. The id column is the true primary key (RPUV=1.00, uniqueness=1.000), but table_classification_id provides better partitioning for typical queries like 'show all key candidates for table X'." } ] } ```