Qdrant icon

Qdrant

Official n8n node to interface with Qdrant - https://qdrant.tech

Actions27

Search → Query Points Groups

AI-generated

Overview

This node operation queries grouped points from a specified collection in Qdrant, a vector search engine. It allows users to perform advanced searches with grouping, filtering, and scoring options on vector data stored in Qdrant collections. This is useful for scenarios where you want to retrieve grouped search results based on vector similarity or other criteria, such as clustering search results by a specific field or analyzing grouped data points.

Use Case Examples

  1. Querying a collection of image feature vectors grouped by category to find similar images within each category.
  2. Performing a grouped search on user behavior vectors to identify clusters of similar user actions.
  3. Retrieving grouped points from a product recommendation vector database to analyze product clusters.

Properties

Name Meaning
Collection Name Name of the collection to query points groups from. It can be selected from a list or specified by name.
Group By Payload field to group the points by, must be a string or number field.
Query The query to perform, which can be a vector, ID, or a complex query object such as recommend, fusion, sample, or formula.
Using Defines which vector name to use for querying.
Filter Conditions to filter points that satisfy specific criteria.
Params Search parameters used when there is no prefetch.
Score Threshold Threshold to return points with scores better than this value.
With Vector Whether to include vectors in the response.
With Payload Whether to include payload data in the response.
Lookup From Location to use for IDs lookup.
Group Size Maximum number of points to return per group.
Limit Maximum number of results to return.
With Lookup Look for points in another collection using the group IDs.
Prefetch Sub-requests to perform first; the query will be performed on the results of these prefetches.
Shard Key Specifies which shards to look for points in; if not specified, all shards are searched.
Consistency Defines read consistency guarantees for the operation, options include Majority, Quorum, and All.
Timeout Overrides the global timeout for this request in seconds.
Request Options Additional request options such as batching, SSL certificate validation, proxy settings, and timeout in milliseconds.

Output

JSON

  • result - The grouped points query results returned from the Qdrant collection, including grouped points data, scores, payloads, and vectors depending on the request options.

Dependencies

  • Qdrant REST API with appropriate credentials

Troubleshooting

  • Ensure the collection name is correct and exists in Qdrant to avoid errors related to missing collections.
  • Verify that the 'Group By' field exists in the payload and is of type string or number to prevent grouping errors.
  • Check the query format and vector names to ensure they match the collection schema.
  • Adjust the score threshold and limit parameters to avoid empty or overly large result sets.
  • If using prefetch or lookup features, ensure the referenced collections and IDs are valid and accessible.
  • Network or authentication issues with the Qdrant REST API may cause request failures; verify credentials and connectivity.

Links

  • Qdrant Documentation - Official documentation for Qdrant vector search engine, useful for understanding query parameters and features.

Discussion