cocoindex-io/cocoindex · error · ValueError

Unexpected schema type: {type(resolved_schema)}

Error message

Unexpected schema type: {type(resolved_schema)}

What it means

_vector_params_from_def expects the resolved schema attached to the vector def to be a recognized schema type (dense or multivector VectorSchema). If the resolved schema is of an unexpected type, this internal-invariant ValueError is raised while building qdrant_models.VectorParams during collection creation.

Source

Thrown at python/cocoindex/connectors/qdrant/_target.py:704

def _vector_params_from_def(
    vector_def: _ResolvedQdrantVectorDef,
) -> qdrant_models.VectorParams:
    """Convert a resolved vector definition to Qdrant VectorParams."""
    resolved_schema = vector_def.schema
    multivector_config = None

    if isinstance(resolved_schema, res_schema.VectorSchema):
        dim = resolved_schema.size
    elif isinstance(resolved_schema, res_schema.MultiVectorSchema):
        dim = resolved_schema.vector_schema.size
        # For multivector, use the specified comparator
        multivector_config = qdrant_models.MultiVectorConfig(
            comparator=_multivector_comparator(vector_def.multivector_comparator)
        )
    else:
        raise ValueError(f"Unexpected schema type: {type(resolved_schema)}")

    return qdrant_models.VectorParams(
        size=dim,
        distance=_distance_from_spec(vector_def.distance),
        multivector_config=multivector_config,
    )


def _sparse_vector_params_from_def(
    sparse_vector_def: _ResolvedQdrantSparseVectorDef,
) -> qdrant_models.SparseVectorParams:
    """Convert a resolved sparse vector definition to Qdrant SparseVectorParams."""
    return qdrant_models.SparseVectorParams(
        modifier=_sparse_modifier_from_spec(sparse_vector_def.modifier)
    )


def _validate_vector_names(names: Collection[str], kind: str) -> None:

View on GitHub (pinned to e84aa99b32)

Solutions

1. Construct schemas only through the public API (QdrantVectorDef with a VectorSchema, resolved via await CollectionSchema.create(...)) instead of instantiating internal resolved types.
2. Check that the VectorSchema passed to QdrantVectorDef is a genuine cocoindex.resources.schema.VectorSchema (or multivector variant), not a look-alike object.
3. Upgrade cocoindex if you rely on internals whose shape changed between versions.

Example fix

// before
schema = CollectionSchema(QdrantVectorDef(schema=custom_obj, distance="cosine"))  # internal ctor
// after
schema = await CollectionSchema.create(vectors=QdrantVectorDef(schema=VectorSchema(dtype=np.float32, size=384), distance="cosine"))

from cocoindex.resources.schema import VectorSchema
assert isinstance(vector_def.schema, VectorSchema), type(vector_def.schema)

isinstance(resolved_schema, (VectorSchema, MultivectorVectorSchema))  # use the public schema classes, not internal resolved types

try:
    schema = await CollectionSchema.create(vectors=def_)
except ValueError as e:
    raise InternalError(f"vector schema unexpected type: {e}") from None

When it happens

Trigger: Reaching collection creation with a QdrantVectorDef whose resolved vector_schema is neither the expected dense nor multivector schema type — typically after constructing/patching a QdrantVectorDef manually or constructing CollectionSchema via the sync __init__ with a hand-built resolved def rather than through the async create() resolver.

Common situations: Subclassing or monkey-patching the schema types; bypassing CollectionSchema.create() and passing arbitrary objects as _vectors; version mismatch between cocoindex and code that builds internal resolved types directly.

Understand the failure class

Background: "This is a bug, please report it": internal invariant violations, unreachable panics, and SNH errors explained — this error's family across 47 libraries.

Related errors

• Unexpected column subkey format: {sub_key!r}, expected to st
• Unsupported record type: {self.record_type}
• Primary key column '{pk}' not found in columns: {list(self.c
• VectorSchemaProvider is required for NumPy ndarray type.
• VectorSchemaProvider only supported for ndarray. Got: {pytho

AI-assisted analysis of cocoindex-io/cocoindex@e84aa99b32 (2026-09-08). Data as JSON: /api/errors/d636f8a54b7b16ab. Report an issue: GitHub.