{
    "table_name": "Iris",
    "primary_key": "Id",
    "candidates": [
        {
            "rank": 0,
            "partition_key": "Id",
            "partition_key_label": "primary_identifier",
            "partition_key_reasoning": "Id is unique per row (150 unique values, uniqueness 1.0, 0% nulls, monotonic 1.0). It identifies each flower sample, so it is the primary identifier. RPUV of 1 would be a poor physical partition for a large table, but for a 150-row dimension the key serves as the identifier and not as a file-level organizer.",
            "clustering_key": null,
            "clustering_key_label": null,
            "clustering_key_reasoning": "Dimension tables do not need internal sorting. The only timestamp, _ingestion_timestamp, is constant (1 unique value) ingestion metadata and is not useful.",
            "world_line": null,
            "session_column": null,
            "confidence": "HIGH",
            "notes": "Small dimension table of 150 rows. Id is the primary key. Species (3 values, RPUV 50, 100% top-3 coverage) is a categorical attribute, not a unique identifier, so it is not used as the key. The _ingestion_timestamp, _batch_id, _source_file and _source_system columns are constant lineage metadata."
        }
    ]
}