Actions26
- Collection Actions
- Payload Actions
- Point Actions
- Search Actions
- Vector Actions
Overview
The node provides an interface to perform batch search queries on a Qdrant collection. Specifically, the "Query Batch Points" operation allows users to send multiple search requests in a single API call to a specified collection. Each search request within the batch can be a vector similarity search, an ID-based search, or a complex query combining different criteria.
This operation is beneficial when you want to efficiently retrieve multiple sets of points from a Qdrant collection without making separate calls for each search. For example, it can be used in recommendation systems, semantic search applications, or any scenario where multiple search queries need to be executed simultaneously against a vector database.
Practical examples:
- Performing several vector similarity searches with different query vectors in one batch.
- Combining ID lookups and vector searches to fetch diverse data points in a single request.
- Running complex filtered queries that combine payload filters and vector similarity in bulk.
Properties
| Name | Meaning |
|---|---|
| Collection Name | The target collection in Qdrant where the batch search will be performed. Can be selected from a list of existing collections or specified by name. |
| Searches | A JSON array representing multiple search requests to execute in batch. Each element can define a vector search, ID search, or a complex query structure. |
| Consistency | Defines the read consistency level for the operation. Options are: Majority, Quorum, or All, which control how many replicas must respond before returning data. |
| Timeout | Overrides the global timeout for this specific request, specified in seconds. Default is 100 seconds. |
Output
The output contains a JSON field with the results of the batch search queries. Each entry corresponds to one search request from the input batch and includes the matching points found in the collection according to the search criteria.
If the node supports binary data output (not explicitly shown here), it would typically represent raw vector data or serialized payloads related to the points retrieved.
Dependencies
- Requires connection to a Qdrant vector search service endpoint.
- Needs an API key or authentication token configured in n8n credentials to authorize requests to the Qdrant REST API.
- The base URL for the Qdrant instance must be set in the node credentials.
Troubleshooting
- Timeout errors: If searches take longer than the specified timeout, increase the timeout value or optimize the search queries.
- Collection not found: Ensure the collection name is correct and exists in the Qdrant instance.
- Invalid search JSON: The "Searches" property must be valid JSON and conform to Qdrant's expected search request format.
- Consistency errors: Using stricter consistency levels (like "All") may cause delays or failures if some replicas are unavailable.
- Authentication failures: Verify that the API key or token is correctly configured and has sufficient permissions.