VectorSchemaProvider only supported for ndarray. Got: {pytho

Error message

VectorSchemaProvider only supported for ndarray. Got: {python_type}

What it means

The Doris target connector can only convert vector column values when the declared Python type maps to numpy ndarray with an explicit vector schema (size/dtype). If a VectorSchemaProvider is attached to a column whose Python type is anything else (list, sequence, etc.), _get_type_mapping rejects it because there is no reliable way to serialize it to Doris's ARRAY<FLOAT>.

Source

Thrown at python/cocoindex/connectors/doris/_target.py:313

        if isinstance(annotation, DorisType):
            return _TypeMapping(annotation.doris_type, annotation.encoder)

    base_type = type_info.base_type

    if base_type in _LEAF_TYPE_MAPPINGS:
        return _LEAF_TYPE_MAPPINGS[base_type]

    if base_type is np.ndarray:
        if vector_schema is None:
            raise ValueError("VectorSchemaProvider is required for NumPy ndarray type.")
        if vector_schema.size <= 0:
            raise ValueError(f"Invalid vector dimension: {vector_schema.size}")
        return _TypeMapping(
            "ARRAY<FLOAT>",
            lambda v: v.tolist() if hasattr(v, "tolist") else list(v),
        )
    elif vector_schema is not None:
        raise ValueError(
            f"VectorSchemaProvider only supported for ndarray. Got: {python_type}"
        )

    if isinstance(
        type_info.variant, (SequenceType, MappingType, RecordType, UnionType, AnyType)
    ):
        return _JSON_MAPPING

    return _JSON_MAPPING


# ============================================================
# Column / Table definitions
# ============================================================


class ColumnDef(NamedTuple):
    type: str  # Doris SQL type

View on GitHub (pinned to e84aa99b32)

Solutions

Change the record field's Python type to numpy.ndarray (np.typing.NDArray[np.float32]) so the vector mapping path is taken
Remove the VectorSchemaProvider override and let the column map as a generic ARRAY type
Convert values to ndarray at write time before declaring rows

Example fix

// before
class Doc(res_schema.Record):
    embedding: list[float] = res_schema.field(vector_schema=VectorSchemaProvider(size=384))
// after
import numpy as np
class Doc(res_schema.Record):
    embedding: np.ndarray = res_schema.field(vector_schema=VectorSchemaProvider(size=384))

Defensive patterns

Strategy: validation

Validate before calling

import numpy as np
assert isinstance(doc.embedding, np.ndarray), f"embedding must be ndarray, got {type(doc.embedding)}"

Type guard

def is_ndarray(v: object) -> bool:
    return isinstance(v, np.ndarray)

Prevention

Type vector fields as np.ndarray in record definitions
Keep VectorSchemaProvider overrides only on ndarray fields
Add a unit test that builds the Doris table schema from your record type

When it happens

Trigger: Declaring a Doris table column with column_overrides containing a res_schema.VectorSchemaProvider for a field whose Python type is not an ndarray-backed vector type (e.g. list[float] or Any), then calling DorisTableSchema.from_class / _columns_from_record_type.

Common situations: Users define their record field as a plain Python list[float] but attach a VectorSchemaProvider expecting Doris VECTOR/ARRAY<FLOAT> mapping; or switch embedding pipelines to return lists instead of numpy arrays while keeping the vector schema override.

Understand the failure class

Background: Type mismatch errors: IllegalArgumentException, TypeError and type guards across 150 open-source libraries — this error's family across 150 libraries.

Related errors

AI-assisted analysis of cocoindex-io/cocoindex@e84aa99b32 (2026-09-08). Data as JSON: /api/errors/12990800d9c2417d. Report an issue: GitHub.