Hi,
I have try to add payload in Qdrant vector store > Collection config.
I have try lot of formation still not success.
Anyone success with this please share some example.
Would you like to show what you were trying to add as a payload and how?
If so, please embed your workflow here.
does this mean it works or it doesn’t?
Fix “still not success”
Do you have any example with this collection config for payload insert.
I have look at document still not working.
You need not to provide the point in the collection config manually. If you are working with a vector store, your input needs to come through as input to the node.
If you would like to just insert some data, use regular Qdrant node.
Which node do you need - Vector Store or regular?
I want to insert both the point and the payload in a single process. While I can select the point and modify it afterward, I don’t think that’s the correct approach for document ingestion.
Metadata alone is not suitable for handling large data. In n8n, when using the Qdrant vector store for embedding, how can I insert the payload during the ingestion process?
Do you have any best practices for this? I assume this should be a standard feature, right?
With vector store node when inserting the document, I don’t believe you can insert information other than the embedding vector and its metadata, but what is the difference between payload and metadata? Everything except the vector is metadata to that vector in a form of a json-like object. Metadata is just a ‘term’ which is used ( primarily by LangChain) to unify the naming across different vector store.
You can almost literally see it in the definition of the payload in qdrant documentation:
What is collection config for?
My understanding this is to setup a collection configuration when the collection doesn’t yet exist, setting the vector dimensionality, distance metrics, datatype as so on.
1 Like