Error in function Node when retrieving from a in-memory vector store

marceltoma · October 12, 2024, 8:04pm

Describe the problem/error/question

I get an error in my function node when I try to use a in-memory vector store using

const vectorStore = await this.getInputConnectionData('ai_vectorStore', 0);

What is the error message (if any)?

Error message in the node

NodeOperationError: Error in sub-node In-Memory Vector Store3 at /usr/local/lib/node_modules/n8n/node_modules/n8n-core/dist/NodeExecuteFunctions.js:1850:19 at processTicksAndRejections (node:internal/process/task_queues:95:5) at async Promise.all (index 0) at Proxy.getInputConnectionData (/usr/local/lib/node_modules/n8n/node_modules/n8n-core/dist/NodeExecuteFunctions.js:1857:19) at Proxy.getInputConnectionData (/usr/local/lib/node_modules/n8n/node_modules/n8n-core/dist/NodeExecuteFunctions.js:2296:24)

Error in the in-memory vector store

context.getWorkflow is not a function
Error details
 Other info
    Item Index 0
    n8n version 1.62.4 (Self Hosted)
    Time 12/10/2024, 16:56:08
    Error cause {}

Please share your workflow

Looks like the error message “context.getWorkflow is not a function” refers to the fact that the in-memory vector store didn’t received the context data.

github.com

n8n-io/n8n/blob/4f27b39b45b58779d363980241e6e5e11b58f5da/packages/@n8n/nodes-langchain/nodes/vector_store/VectorStoreInMemory/VectorStoreInMemory.node.ts#L45


      
          			type: 'string',
          			default: 'vector_store_key',
          			description:
          				'The key to use to store the vector memory in the workflow data. The key will be prefixed with the workflow ID to avoid collisions.',
          		},
          	],
          	insertFields,
          	loadFields: [],
          	retrieveFields: [],
          	async getVectorStoreClient(context, _filter, embeddings, itemIndex) {
          		const workflowId = context.getWorkflow().id;
          		const memoryKey = context.getNodeParameter('memoryKey', itemIndex) as string;
          		const vectorStoreSingleton = MemoryVectorStoreManager.getInstance(embeddings);
          
          		return await vectorStoreSingleton.getVectorStore(`${workflowId}__${memoryKey}`);
          	},
          	async populateVectorStore(context, embeddings, documents, itemIndex) {
          		const memoryKey = context.getNodeParameter('memoryKey', itemIndex) as string;
          		const clearStore = context.getNodeParameter('clearStore', itemIndex) as boolean;
          		const workflowId = context.getWorkflow().id;
          		const vectorStoreInstance = MemoryVectorStoreManager.getInstance(embeddings);

Information on your n8n setup

n8n version: 1.62.4
Database (default: SQLite):
n8n EXECUTIONS_PROCESS setting (default: own, main):
Running n8n via (Docker, npm, n8n cloud, desktop app): Docker
Operating system: Windows 11

marceltoma · October 14, 2024, 6:37pm

https://github.com/n8n-io/n8n/issues/11256

Jon · October 15, 2024, 10:31am

Hey @marceltoma,

Do you have a full example workflow and data we can use to test this? I suspect it is likely something missing in the node but having an example we can just run to test will be very helpful.

marceltoma · October 15, 2024, 11:36am

Here is an example

system · January 13, 2025, 11:37am

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.