{
    "table_name": "Indian Startups _ Funding  Investors Data April 2022",
    "primary_key": "Company Name",
    "candidates": [
        {
            "rank": 0,
            "partition_key": "Company Name",
            "partition_key_label": "primary_identifier",
            "partition_key_reasoning": "Company Name has 94 unique values across 95 rows (uniqueness 0.989, RPUV 1.01), so it identifies one company per row. Only 'Securden' appears twice, and the two rows are an exact duplicate apart from the Amount formatting ($11.700,000 vs $11,700,000). Top-3 coverage is only 4.21%.",
            "clustering_key": null,
            "clustering_key_label": null,
            "clustering_key_reasoning": "Dimension tables do not need internal sorting. There is no time axis; the only datetime column is the constant ingestion timestamp.",
            "world_line": null,
            "session_column": null,
            "confidence": "MEDIUM",
            "notes": "Company Name is the natural identifier for this per-company funding profile. It is nearly unique, but the duplicated Securden row means the key should be deduplicated before use as a strict primary key. The table is small (95 rows), so the partition key could also be null. Other columns (Location, Sector, Stage) are attributes, not entity identifiers. The ingestion metadata columns are constants and unusable as keys."
        }
    ]
}