PGVectorStore faild with AzureAIEmbeddingsModel (Document objects could not be iterable) #93

New Issue

2026-02-16T05:16:32-05:00

yindo commented

2026-02-16 05:16:32 -05:00

Originally created by @pccrdotnet on GitHub (Sep 20, 2025).

Originally assigned to: @averikitsch on GitHub.

At the moment, when I try to create documents using the method store.aadd_documents() with AzureAIEmbeddingsModel as embedding_service it shows an error about an iterable problem (TypeError: 'async_generator' object is not iterable). I think the problem could be when the async method tries to read the metadata object to interact for the insertion operation into the DB, next this is the exception:

File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/vectorstores.py", line 218, in aadd_documents
return await self._engine._run_as_async(
File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/engine.py", line 121, in _run_as_async
return await asyncio.wrap_future(
File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/async_vectorstore.py", line 397, in aadd_documents
ids = await self.aadd_texts(texts, metadatas=metadatas, ids=ids, **kwargs)
File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/async_vectorstore.py", line 377, in aadd_texts
ids = await self.aadd_embeddings(
File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/async_vectorstore.py", line 286, in aadd_embeddings
for id, content, embedding, metadata in zip(ids, texts, embeddings, metadatas):
TypeError: 'async_generator' object is not iterable

Example of the code with the error:

import asyncio
from langchain_postgres.v2.engine import Column
from langchain_postgres import PGEngine, PGVectorStore
from langchain_core.embeddings import DeterministicFakeEmbedding
from langchain_core.documents import Document
from langchain_azure_ai.embeddings import AzureAIEmbeddingsModel

async def StartAgent() -> None:
    # Configuration
    CONNECTION_STRING = "postgresql+psycopg://postgres:XXXXXXXXXX@localhost:5432/TestDb3"
    VECTOR_SIZE = 1536
    TABLE_NAME = "my_doc_collection2"

    azKey = "XXXXXXXXXXX"
    azEndPointEmbedding = "https://XXXXXXXXX.openai.azure.com/openai/deployments/text-embedding-ada-002/"
    azModelEmbedding = "text-embedding-ada-002"

    # Initialize engine and embedding
    engine = PGEngine.from_connection_string(url=CONNECTION_STRING)

    embedding = AzureAIEmbeddingsModel(
            endpoint=azEndPointEmbedding,
            credential=azKey,
            model_name=azModelEmbedding
        )

    #Define metadata columns when creating the table
    # engine.init_vectorstore_table(
    #     table_name=TABLE_NAME,
    #     vector_size=VECTOR_SIZE,
    #     metadata_columns=[
    #         Column(name="country", data_type="text"),
    #         Column(name="city", data_type="text"),
    #         Column(name="address", data_type="text"),
    #         # Add more columns as needed
    #     ]
    # )

    # Create the vector store
    store = await PGVectorStore.create(
        engine=engine,
        table_name=TABLE_NAME,
        embedding_service=embedding,
        metadata_columns=["country", "city", "address"]  # Must match the columns defined above
    )

    # Add documents with metadata
    docs = [
        Document(
            page_content="Apples and oranges",
            metadata={"country": "USA", "city": "New York", "address": "123 Main St"}
        ),
        Document(
            page_content="Cars and airplanes",
            metadata={"country": "France", "city": "Paris", "address": "456 Rue de la Paix"}
        ),
        Document(
            page_content="Train",
            metadata={"country": "Japan", "city": "Tokyo", "address": "789 Shinjuku Ave"}
        )
    ]

    await store.aadd_documents(documents=docs)


if __name__ == "__main__":    
    asyncio.run(StartAgent())

Libraries:
langchain 0.3.27
langchain-azure-ai 0.1.5
langchain-cohere 0.4.6
langchain-community 0.3.27
langchain-core 0.3.76
langchain-experimental 0.3.4
langchain-openai 0.3.33
langchain-postgres 0.0.15
langchain-text-splitters 0.3.9
langgraph 0.6.7
langgraph-checkpoint 2.1.1
langgraph-prebuilt 0.6.4
langgraph-sdk 0.2.6
langsmith 0.4.14

Note: This code works fine when the embedding service changes to DeterministicFakeEmbedding.

Originally created by @pccrdotnet on GitHub (Sep 20, 2025). Originally assigned to: @averikitsch on GitHub. At the moment, when I try to create documents using the method store.aadd_documents() with AzureAIEmbeddingsModel as embedding_service it shows an error about an iterable problem (TypeError: 'async_generator' object is not iterable). I think the problem could be when the async method tries to read the metadata object to interact for the insertion operation into the DB, next this is the exception: File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/vectorstores.py", line 218, in aadd_documents return await self._engine._run_as_async( File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/engine.py", line 121, in _run_as_async return await asyncio.wrap_future( File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/async_vectorstore.py", line 397, in aadd_documents ids = await self.aadd_texts(texts, metadatas=metadatas, ids=ids, **kwargs) File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/async_vectorstore.py", line 377, in aadd_texts ids = await self.aadd_embeddings( File "/home/pablo/projects/RAGPrototype/.venv/lib/python3.10/site-packages/langchain_postgres/v2/async_vectorstore.py", line 286, in aadd_embeddings for id, content, embedding, metadata in zip(ids, texts, embeddings, metadatas): TypeError: 'async_generator' object is not iterable Example of the code with the error: ``` import asyncio from langchain_postgres.v2.engine import Column from langchain_postgres import PGEngine, PGVectorStore from langchain_core.embeddings import DeterministicFakeEmbedding from langchain_core.documents import Document from langchain_azure_ai.embeddings import AzureAIEmbeddingsModel async def StartAgent() -> None: # Configuration CONNECTION_STRING = "postgresql+psycopg://postgres:XXXXXXXXXX@localhost:5432/TestDb3" VECTOR_SIZE = 1536 TABLE_NAME = "my_doc_collection2" azKey = "XXXXXXXXXXX" azEndPointEmbedding = "https://XXXXXXXXX.openai.azure.com/openai/deployments/text-embedding-ada-002/" azModelEmbedding = "text-embedding-ada-002" # Initialize engine and embedding engine = PGEngine.from_connection_string(url=CONNECTION_STRING) embedding = AzureAIEmbeddingsModel( endpoint=azEndPointEmbedding, credential=azKey, model_name=azModelEmbedding ) #Define metadata columns when creating the table # engine.init_vectorstore_table( # table_name=TABLE_NAME, # vector_size=VECTOR_SIZE, # metadata_columns=[ # Column(name="country", data_type="text"), # Column(name="city", data_type="text"), # Column(name="address", data_type="text"), # # Add more columns as needed # ] # ) # Create the vector store store = await PGVectorStore.create( engine=engine, table_name=TABLE_NAME, embedding_service=embedding, metadata_columns=["country", "city", "address"] # Must match the columns defined above ) # Add documents with metadata docs = [ Document( page_content="Apples and oranges", metadata={"country": "USA", "city": "New York", "address": "123 Main St"} ), Document( page_content="Cars and airplanes", metadata={"country": "France", "city": "Paris", "address": "456 Rue de la Paix"} ), Document( page_content="Train", metadata={"country": "Japan", "city": "Tokyo", "address": "789 Shinjuku Ave"} ) ] await store.aadd_documents(documents=docs) if __name__ == "__main__": asyncio.run(StartAgent()) ``` Libraries: langchain 0.3.27 langchain-azure-ai 0.1.5 langchain-cohere 0.4.6 langchain-community 0.3.27 langchain-core 0.3.76 langchain-experimental 0.3.4 langchain-openai 0.3.33 langchain-postgres 0.0.15 langchain-text-splitters 0.3.9 langgraph 0.6.7 langgraph-checkpoint 2.1.1 langgraph-prebuilt 0.6.4 langgraph-sdk 0.2.6 langsmith 0.4.14 Note: This code works fine when the embedding service changes to DeterministicFakeEmbedding.

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: langchain-ai/langchain-postgres#93