lancedb/lancedb · error · ValueError

Unknown index type

Error message

Unknown index type {index_type}

What it means

The legacy create_index path dispatches on index_type and only supports the recognized vector index variants (IVF_PQ, IVF_HNSW_PQ, etc.); anything else raises ValueError('Unknown index type {index_type}').

Solutions

1. Use one of the accepted index_type strings (e.g. "IVF_PQ", "IVF_HNSW_PQ", "IVF_HNSW_SQ").
2. Prefer the config-based API: table.create_index(column, config=IvfPq(...)) instead of index_type strings.
3. For scalar indexes use BTree/Bitmap/LabelList configs with create_index(column=...).
4. Check the signature/docs of your installed lancedb version; index_type names changed over releases.

Example fix

// before
table.create_index(num_partitions=16, index_type="IVFPQ")
// after
from lancedb.index import IvfPq
table.create_index(metric="L2", vector_column_name="vector",
                   index_type="IVF_PQ")
# or config-based:
table.create_index(config=IvfPq(num_partitions=16))

ALLOWED = {"IVF_PQ", "IVF_HNSW_PQ", "IVF_HNSW_SQ"}
if index_type not in ALLOWED:
    raise ValueError(f"index_type must be one of {ALLOWED}")

try:
    table.create_index(index_type=index_type, ...)
except ValueError as e:
    if "Unknown index type" in str(e):
        logger.error(f"Unsupported index_type: {index_type}")

When it happens

Trigger: Calling table.create_index(metric=..., vector_column_name=..., index_type=X) where X is not one of the accepted strings (e.g. 'IVFPQ', 'hnsw', lowercase 'ivf_pq', or a scalar type like 'BTREE' on this legacy path).

Common situations: Copy-pasted index_type from another library (FAISS/pgvector style); case mistakes; trying to create a scalar index (BTREE/BITMAP/LABEL_LIST) through the legacy vector API instead of the config-based create_index.

Understand the failure class

Background: Invalid enum value errors: "Unknown type", "Invalid scope", "must be one of" — when a string is not on the library's allowed list — this error's family across 23 libraries.

Related errors

• Cannot select column
• config must be an instance of IvfSq, IvfPq, IvfRq, HnswPq…
• Either data or schema is required for server-side…
• fragments and fragment_ids cannot both be set
• Invalid format

AI-assisted analysis of lancedb/lancedb@c7b051aff7 (2026-09-08). Data as JSON: /api/errors/b153f6f9a387c9b5. Report an issue: GitHub.

Appendix: source

Thrown at python/python/lancedb/table.py:3326

                max_iterations=max_iterations,
                sample_rate=sample_rate,
                m=m,
                ef_construction=ef_construction,
                target_partition_size=target_partition_size,
                accelerator=accelerator,
            )
        elif index_type == "IVF_HNSW_FLAT":
            return HnswFlat(
                distance_type=metric,
                num_partitions=num_partitions,
                max_iterations=max_iterations,
                sample_rate=sample_rate,
                m=m,
                ef_construction=ef_construction,
                target_partition_size=target_partition_size,
            )
        else:
            raise ValueError(f"Unknown index type {index_type}")

    def drop_index(self, name: str) -> None:
        """
        Drops an index from the table

        Parameters
        ----------
        name: str
            The name of the index to drop
        """
        return LOOP.run(self._table.drop_index(name))

    def prewarm_index(self, name: str) -> None:
        """
        Prewarm an index in the table.

        This is a hint to the database that the index will be accessed in the
        future and should be loaded into memory if possible.  This can reduce

View on GitHub (pinned to c7b051aff7)