[BUG]: VectorDB Retriever seems to not use id_property_external #448
Description
Before You Report a Bug, Please Confirm You Have Done The Following...
- I have updated to the latest version of the packages.
- I have searched for both existing issues and closed issues and found none that matched my issue.
neo4j-graphrag-python's version
1.10.1
Python version
3.10.7
Operating System
Windows & Macos
Dependencies
Using the QdrantNeo4jRetriever , it looks like no matter what I put in id_property_external , it always use the same id from my vector store:
i.e
id_property_external="uid",
id_property_neo4j="uid",with top k of 1, gives me:
{'match_params': [['5383723d-06fb-414e-af62-d81e03db2852', 0.7332783]]], 'id_property': 'uid'}note the value and format of the uid 5383...
id_property_external="test-whatever", # but same if I use another real attribute
id_property_neo4j="uid",(or any real metadata that exists in my vector db instead of "test-whatever")
Also gives me:
{'match_params': [['5383723d-06fb-414e-af62-d81e03db2852', 0.7332783]]], 'id_property': 'uid'}Same results, id used is the same 5383...
When looking at the code, not sure I find the issue as I see you do update the parameter payload with this external property id: https://github.com/neo4j/neo4j-graphrag-python/blob/main/src/neo4j_graphrag/retrievers/external/qdrant/qdrant.py#L220
More in slack:
https://neo4j.slack.com/archives/C092146TSR4/p1764265357649359
Code Snippet shared below in reproducible example
Reproducible example
import os
import logging
from qdrant_client import QdrantClient
from dotenv import load_dotenv
from neo4j import GraphDatabase
from neo4j_graphrag.retrievers import QdrantNeo4jRetriever
from neo4j_graphrag.llm import OpenAILLM
from neo4j_graphrag.generation import GraphRAG
from neo4j_graphrag.embeddings import Embedder
from fastembed import TextEmbedding
from typing import Any
logging.basicConfig(
level=logging.DEBUG,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)
class FastEmbedEmbedder(Embedder):
def __init__(self, model_name: str = "nomic-ai/nomic-embed-text-v1", **kwargs: Any):
super().__init__(**kwargs)
self.model = TextEmbedding(model_name=model_name)
def embed_query(self, text: str) -> list[float]:
embeddings = list(self.model.embed([text]))
result = embeddings[0].tolist()
return result
def main():
load_dotenv()
QDRANT_URL = os.getenv('QDRANT_URL')
QDRANT_API_KEY = os.getenv('QDRANT_API_KEY')
QDRANT_COLLECTION_NAME = os.getenv("QDRANT_COLLECTION_NAME")
QDRANT_PORT = os.getenv("QDRANT_PORT")
NEO4J_URI = os.getenv("NEO4J_URI")
NEO4J_USERNAME = os.getenv("NEO4J_USERNAME")
NEO4J_PASSWORD = os.getenv("NEO4J_PASSWORD")
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
qdrant_client = QdrantClient(
url=QDRANT_URL,
api_key=QDRANT_API_KEY,
port=QDRANT_PORT,
verify=False,
)
driver = GraphDatabase.driver(NEO4J_URI, auth=(NEO4J_USERNAME, NEO4J_PASSWORD))
embedder = FastEmbedEmbedder(model_name="nomic-ai/nomic-embed-text-v1")
retriever = QdrantNeo4jRetriever(
driver=driver,
client=qdrant_client,
collection_name=QDRANT_COLLECTION_NAME,
using="nomic-ai/nomic-embed-text-v1",
id_property_external="test-whatever",
id_property_neo4j="uid",
embedder=embedder,
retrieval_query="MATCH (node:Section) WHERE node.uid=match_id_value RETURN node LIMIT 5"
)
query_text = "How was Gas detection tests performed?"
ret = retriever.search(query_text=query_text, top_k=5)
print(ret)Relevant Log Output
2025-11-28 08:24:00,163 - httpx - INFO - HTTP Request: POST https://qdrant-poc.gliadatalake-poc.corp.aws.novonordisk.com/collections/rag_poc_hybrid/points/query "HTTP/1.1 200 OK"
2025-11-28 08:24:00,163 - httpcore.http11 - DEBUG - receive_response_body.started request=<Request [b'POST']>
2025-11-28 08:24:00,165 - httpcore.http11 - DEBUG - receive_response_body.complete
2025-11-28 08:24:00,166 - httpcore.http11 - DEBUG - response_closed.started
2025-11-28 08:24:00,166 - httpcore.http11 - DEBUG - response_closed.complete
2025-11-28 08:24:00,168 - neo4j_graphrag.retrievers.external.qdrant.qdrant - DEBUG - Qdrant Store Cypher parameters: {'match_params': [['5383723d-06fb-414e-af62-d81e03db2852', 0.7332783], ['70155aa7-522a-4b4c-be84-d823c26ba985', 0.71326256], ['81ea44be-61da-423e-a75c-5cdc0d62ada2', 0.70934486], ['51575ba0-ae90-4bfb-a3f0-0ce65198fa40', 0.7056217], ['5bf30796-3091-47ac-acd2-07e4faf57ab7', 0.7024784]], 'id_property': 'uid'}
2025-11-28 08:24:00,168 - neo4j_graphrag.retrievers.external.qdrant.qdrant - DEBUG - Qdrant Store Cypher query: UNWIND $match_params AS match_param WITH match_param[0] AS match_id_value, match_param[1]
AS score MATCH (node) WHERE node[$id_property] = match_id_value MATCH (node:Section) WHERE node.uid=match_id_value RETURN node LIMIT 5
Expected Result
The id value in the generated query below to change based on the id_property_external used
{'match_params': [['5383723d-06fb-414e-af62-d81e03db2852', 0.7332783], ['70155aa7-522a-4b4c-be84-d823c26ba985', 0.71326256], ['81ea44be-61da-423e-a75c-5cdc0d62ada2', 0.70934486], ['51575ba0-ae90-4bfb-a3f0-0ce65198fa40', 0.7056217], ['5bf30796-3091-47ac-acd2-07e4faf57ab7', 0.7024784]], 'id_property': 'uid'}
What happened instead?
no matter what property I use in id_property_external, it always use the same IDs
Additional Info
No response