```json { "table_name": "dataset_lifecycle", "primary_key": "batch_id", "candidates": [ { "rank": 0, "partition_key": "dataset_id", "partition_key_label": "E", "partition_key_reasoning": "RPUV=2.00 indicates each dataset has ~2 batches on average, making it a good partition key. Uniqueness=0.500 with 99 distinct datasets out of 198 rows. Top 3 coverage=24.75% is well below 50%. Semantically, dataset_id is the primary entity whose lifecycle we are tracking through multiple batch submissions.", "clustering_key": "submitted_at", "clustering_key_label": "S", "clustering_key_reasoning": "submitted_at is the event timestamp with uniqueness=1.000 (all unique), making it ideal for ordering events within each dataset's timeline. This is the time axis identified by the classifier. Ordering by submission time gives chronological progression of each dataset's lifecycle events.", "world_line": "dataset_id", "session_column": null, "confidence": "HIGH", "notes": "Primary perspective: track each dataset's lifecycle through its batch submissions over time. dataset_id groups related events, submitted_at orders them chronologically." }, { "rank": 1, "partition_key": "batch_id", "partition_key_label": "E", "partition_key_reasoning": "RPUV=1.00 with uniqueness=1.000 (198 unique batches for 198 rows). Each batch_id appears exactly once, making it a valid but less useful partition key. Top 3 coverage=1.52% is excellent. Semantically, batch_id represents individual processing runs.", "clustering_key": "submitted_at", "clustering_key_label": "S", "clustering_key_reasoning": "submitted_at provides temporal ordering even when partitioned by batch_id, though with RPUV=1.00 each partition would contain only one row, making clustering less meaningful.", "world_line": "batch_id", "session_column": null, "confidence": "MEDIUM", "notes": "Alternative perspective: partition by batch_id for batch-centric analysis. However, with RPUV=1.00, each partition contains only one row, making this less efficient than dataset_id partitioning. Useful if queries primarily filter by specific batch IDs." } ] } ```