microsoft/semantic-kernel · error · ServiceResourceNotFoundError

No match found

Error message

No match found

What it means

Raised by PostgresMemoryStore.get_nearest_match() when the underlying get_nearest_matches(limit=1) returns an empty list. Unlike the collection-missing case, the collection exists but no row cleared the min_relevance_score threshold. ServiceResourceNotFoundError with literal message 'No match found'.

Source

Thrown at python/semantic_kernel/connectors/memory_stores/postgres/postgres_memory_store.py:475

        Args:
            collection_name: The name of the collection to get the nearest match from.
            embedding: The embedding to find the nearest match to.
            min_relevance_score: The minimum relevance score of the match. (default: {0.0})
            with_embedding: Whether to include the embedding in the result. (default: {False})

        Returns:
            Tuple[MemoryRecord, float]: The record and the relevance score.
        """
        results = await self.get_nearest_matches(
            collection_name=collection_name,
            embedding=embedding,
            limit=1,
            min_relevance_score=min_relevance_score,
            with_embeddings=with_embedding,
        )
        if len(results) == 0:
            raise ServiceResourceNotFoundError("No match found")
        return results[0]

    async def __does_collection_exist(self, cur: Cursor, collection_name: str) -> bool:
        results = await self.__get_collections(cur)
        return collection_name in results

    async def __get_collections(self, cur: Cursor) -> list[str]:
        cur.execute(
            """
            SELECT table_name
            FROM information_schema.tables
            WHERE table_schema = %s
            """,
            (self._schema,),
        )
        return [row[0] for row in cur.fetchall()]

    def _check_dimensionality(self, dimension_num):

View on GitHub (pinned to c028a0c7dc)

Solutions

1. Lower or remove min_relevance_score (default 0.0) and retry.
2. Use get_nearest_matches(limit, ...) directly and handle an empty list instead of relying on get_nearest_match() to throw.
3. Verify the collection has data and that query embeddings come from the same model/dimensionality.
4. Catch ServiceResourceNotFoundError and degrade to a fallback response.

Example fix

// before
best = await store.get_nearest_match('mycol', emb, min_relevance_score=0.9)
// after
matches = await store.get_nearest_matches('mycol', emb, limit=1, min_relevance_score=0.0)
best = matches[0] if matches else None

matches = await store.get_nearest_matches(collection_name, embedding, limit=1, min_relevance_score=min_relevance_score)
best = matches[0] if matches else None

from semantic_kernel.exceptions import ServiceResourceNotFoundError
try:
    best = await store.get_nearest_match(collection_name, embedding, min_relevance_score)
except ServiceResourceNotFoundError:
    best = None

When it happens

Trigger: Calling `await store.get_nearest_match(collection_name, embedding, min_relevance_score=...)` on an existing collection where either the table is empty or every candidate scores below min_relevance_score.

Common situations: min_relevance_score set too high (e.g. > best cosine similarity); empty/freshly created collection; embedding from a different model than stored vectors (scores near 0); querying the wrong collection.

Related errors

• Key not found
• Failed to create Postgres settings.
• Collection '{collection_name}' does not exist
• Upsert failed
• Dimensionality of {dimension_num} exceeds the maximum allowe

AI-assisted analysis of microsoft/semantic-kernel@c028a0c7dc (2026-08-13). Data as JSON: /api/errors/8738d30adbedf0e9. Report an issue: GitHub.