Actions26
- Collection Actions
- Payload Actions
- Point Actions
- Search Actions
- Vector Actions
Overview
This node integrates with Qdrant, a vector search engine, to manage points within collections. Specifically, the "Upsert Points" operation for the "Point" resource allows users to insert new points or update existing ones in a specified collection. Each point consists of an ID, a vector (numerical array representing features), and optional payload data (metadata).
Common scenarios include:
- Adding new data points to a vector database for similarity search.
- Updating existing points with new vectors or metadata.
- Maintaining up-to-date vector representations for machine learning or recommendation systems.
Practical example:
- A user has a collection of product embeddings and wants to add new products or update existing product vectors and their attributes like color or age.
Properties
| Name | Meaning |
|---|---|
| Collection Name | The target collection where points will be upserted. Can be selected from a list or specified by name. |
| Points | JSON array of points to upsert. Each point must have: • id: unique identifier • vector: numerical array representing the point's features • Optional payload: object containing metadata such as attributes or tags. |
| Wait | Boolean flag indicating whether the operation should wait until changes are fully applied before returning. |
| Ordering | Defines ordering guarantees for the operation. Options: • Weak • Medium • Strong |
Output
The node outputs JSON data representing the result of the upsert operation. This typically includes confirmation of successful insertion or update of points. The exact structure depends on the Qdrant API response but generally confirms which points were processed.
No binary data output is involved in this operation.
Dependencies
- Requires connection to a Qdrant instance via REST API.
- Needs an API authentication token or key configured in n8n credentials to authorize requests.
- The base URL of the Qdrant service must be set in the node credentials.
Troubleshooting
Common issues:
- Invalid or missing collection name: Ensure the collection exists or is correctly named.
- Malformed points JSON: Verify that each point has a valid
idandvectorarray. - Authentication errors: Check that the API key/token is valid and properly configured.
- Network connectivity problems: Confirm that the Qdrant service URL is reachable.
Error messages:
- "Collection not found": The specified collection does not exist; create it first or correct the name.
- "Invalid vector format": Vectors must be arrays of numbers; check the input JSON.
- "Unauthorized": Authentication failed; verify credentials.
- Timeouts or network errors: Check network settings and Qdrant server status.