Qdrant

Qdrant source connector

Description

Qdrant is a high-performance vector search engine and vector database.

This connector can be used to read data from a Qdrant collection.

Options

name	type	required	default value
collection_name	string	yes	-
schema	config	yes	-
host	string	no	localhost
port	int	no	6334
api_key	string	no	-
use_tls	int	no	false
common-options		no	-

collection_name [string]

The name of the Qdrant collection to read data from.

schema [config]

The schema of the table to read data into.

Eg:

schema = {
  fields {
    age = int
    address = string
    some_vector = float_vector
  }
}

Each entry in Qdrant is called a point.

The float_vector type columns are read from the vectors of each point, others are read from the JSON payload associated with the point.

If a column is marked as primary key, the ID of the Qdrant point is written into it. It can be of type "string" or "int". Since Qdrant only allows positive integers and UUIDs as point IDs.

If the collection was created with a single default/unnamed vector, use default_vector as the vector name.

schema = {
  fields {
    age = int
    address = string
    default_vector = float_vector
  }
}

The ID of the point in Qdrant will be written into the column which is marked as the primary key. It can be of type int or string.

host [string]

The host name of the Qdrant instance. Defaults to "localhost".

port [int]

The gRPC port of the Qdrant instance.

api_key [string]

The API key to use for authentication if set.

use_tls [bool]

Whether to use TLS(SSL) connection. Required if using Qdrant cloud(https).

common options

Source plugin common parameters, please refer to Source Common Options for details.