run-llama/llama_index · error · NotImplementedError

Schema not implemented for SimplePropertyGraphStore.

Error message

Schema not implemented for SimplePropertyGraphStore.

What it means

SimplePropertyGraphStore.get_schema() unconditionally raises NotImplementedError — the class comment marks these as intentionally unimplemented methods. This in-memory property-graph store (nodes + weighted edges in a LabelledPropertyGraph model) supports persistence and triple-style accessors but not the schema introspection that graph-database backends provide.

Source

Thrown at llama-index-core/llama_index/core/graph_stores/simple_labelled.py:235

        data["nodes"] = {}

        # load the graph
        graph = LabelledPropertyGraph.model_validate(data)

        # add the node back
        graph.nodes = kg_nodes

        return cls(graph)

    def to_dict(self) -> dict:
        """Convert to dict."""
        return self.graph.model_dump()

    # NOTE: Unimplemented methods for SimplePropertyGraphStore

    def get_schema(self, refresh: bool = False) -> str:
        """Get the schema of the graph store."""
        raise NotImplementedError(
            "Schema not implemented for SimplePropertyGraphStore."
        )

    def structured_query(
        self, query: str, param_map: Optional[Dict[str, Any]] = None
    ) -> Any:
        """Query the graph store with statement and parameters."""
        raise NotImplementedError(
            "Structured query not implemented for SimplePropertyGraphStore."
        )

    def vector_query(
        self, query: VectorStoreQuery, **kwargs: Any
    ) -> Tuple[List[LabelledNode], List[float]]:
        """Query the graph store with a vector store query."""
        raise NotImplementedError(
            "Vector query not implemented for SimplePropertyGraphStore."
        )

View on GitHub (pinned to afd0fef371)

Solutions

  1. Use a schema-aware property graph backend (Neo4j, Neptune, Kuzu, TigerGraph integrations) when you need get_schema().
  2. Skip schema retrieval for this store: catch NotImplementedError or feature-check before calling, and build prompts from your own static schema description.
  3. Explore the graph via supported APIs (get/upsers of nodes and edges, triple accessors) instead.

Example fix

# before
schema = store.get_schema(refresh=True)  # SimplePropertyGraphStore

# after
try:
    schema = store.get_schema(refresh=True)
except NotImplementedError:
    schema = MY_STATIC_SCHEMA_TEXT  # describe entities/relations yourself
Defensive patterns

Strategy: fallback

Validate before calling

from llama_index.core.graph_stores.simple_labelled import SimplePropertyGraphStore

schema = None if isinstance(store, SimplePropertyGraphStore) else store.get_schema(refresh=True)

Type guard

def supports_schema(store) -> bool:
    return not isinstance(store, SimplePropertyGraphStore)

Try / catch

try:
    schema = store.get_schema(refresh=True)
except NotImplementedError:
    schema = STATIC_SCHEMA_DESCRIPTION  # your own entity/relation summary

Prevention

When it happens

Trigger: Calling get_schema() on a SimplePropertyGraphStore — most often from PropertyGraphIndex flows that build a schema-aware prompt, or generic code that was originally written against Neo4jPropertyGraphStore.

Common situations: Using SimplePropertyGraphStore as a zero-dependency default for PropertyGraphIndex and then enabling schema-based query generation; swapping the store backend via config without disabling schema prompts; agents that call get_schema() to orient themselves.

Related errors

AI-assisted analysis of run-llama/llama_index@afd0fef371 (2026-08-15). Data as JSON: /api/errors/fcb387d3447a42c4. Report an issue: GitHub.