microsoft/semantic-kernel · error · VectorStoreOperationException

Vector is required for search.

Error message

Vector is required for search.

What it means

Raised by MongoDBAtlasCollection._inner_search (VectorStoreOperationException) as the fallback when search_type is neither VECTOR nor KEYWORD_HYBRID — the only two modes this connector implements. Despite the message ('Vector is required for search.'), the real cause is an unsupported search type (e.g. a text-only KEYWORD search or a future/new SearchType value).

Source

Thrown at python/semantic_kernel/connectors/mongodb.py:354

    @override
    async def ensure_collection_deleted(self, **kwargs) -> None:
        await self._get_database().drop_collection(self.collection_name, **kwargs)

    @override
    async def _inner_search(
        self,
        search_type: SearchType,
        options: VectorSearchOptions,
        values: Any | None = None,
        vector: Sequence[float | int] | None = None,
        **kwargs: Any,
    ) -> KernelSearchResults[VectorSearchResult[TModel]]:
        if search_type == SearchType.VECTOR:
            return await self._inner_vector_search(options, values, vector, **kwargs)
        if search_type == SearchType.KEYWORD_HYBRID:
            return await self._inner_keyword_hybrid_search(options, values, vector, **kwargs)
        raise VectorStoreOperationException("Vector is required for search.")

    async def _inner_vector_search(
        self,
        options: VectorSearchOptions,
        values: Any | None = None,
        vector: Sequence[float | int] | None = None,
        **kwargs: Any,
    ) -> KernelSearchResults[VectorSearchResult[TModel]]:
        collection = self._get_collection()
        vector_field = self.definition.try_get_vector_field(options.vector_property_name)
        if not vector_field:
            raise VectorStoreModelException(
                f"Vector field '{options.vector_property_name}' not found in the data model definition."
            )
        if not vector:
            vector = await self._generate_vector_from_values(values, options)
        vector_search_query: dict[str, Any] = {
            "limit": options.top + options.skip,

View on GitHub (pinned to c028a0c7dc)

Solutions

1. Use SearchType.VECTOR (pure vector) or SearchType.KEYWORD_HYBRID for this collection.
2. For text-only keyword search, use a dedicated text-search collection/store instead of the Atlas vector collection.
3. Validate search_type against {VECTOR, KEYWORD_HYBRID} before calling search.

Example fix

// before
await collection.search(search_type=SearchType.KEYWORD, vector=emb)
// after
await collection.search(search_type=SearchType.VECTOR, vector=emb)

from semantic_kernel.data.vector import SearchType
allowed = {SearchType.VECTOR, SearchType.KEYWORD_HYBRID}
if search_type not in allowed:
    raise ValueError(f'MongoDBAtlasCollection supports only {allowed}')
await collection.search(search_type=search_type, vector=emb)

from semantic_kernel.data.vector import SearchType

def is_supported_search(st) -> bool:
    return st in {SearchType.VECTOR, SearchType.KEYWORD_HYBRID}

from semantic_kernel.exceptions import VectorStoreOperationException
try:
    await collection.search(search_type=search_type, vector=emb)
except VectorStoreOperationException as e:
    if 'required for search' in str(e):
        await collection.search(search_type=SearchType.VECTOR, vector=emb)
    else:
        raise

When it happens

Trigger: Calling `collection.search(search_type=SearchType.KEYWORD, ...)` or passing a custom/raw search_type string. SearchType currently has only VECTOR and KEYWORD_HYBRID, so this is typically a new enum member or an externally-supplied value.

Common situations: Using generic cross-connector search code that passes SearchType.KEYWORD (text search); upgrading SK and using a new SearchType this connector hasn't mapped; misrouting a text-only query to the vector collection.

Related errors

• Vector field '{options.vector_property_name}' not found in t
• Failed to search the collection.
• Vector field '{options.vector_property_name}' not found in t
• Failed to search the collection.
• Index kind '{field.index_kind}' is not supported by Azure Co

AI-assisted analysis of microsoft/semantic-kernel@c028a0c7dc (2026-08-13). Data as JSON: /api/errors/0d67e87930345842. Report an issue: GitHub.