Actions35
- Genie Actions
- Databricks SQL Actions
- Unity Catalog Actions
- Model Serving Actions
- Files Actions
- Vector Search Actions
Overview
The "Create Index" operation in the Vector Search resource allows users to create a new vector search index on a specified serving endpoint. This index is designed to store and manage vector data, enabling efficient similarity searches or nearest neighbor queries based on vector representations.
Common scenarios for this node include:
- Setting up a vector search index for machine learning applications such as recommendation systems, image or text similarity search.
- Preparing an index to store embeddings generated from unstructured data like documents, images, or audio.
- Initializing vector indexes before inserting or querying vector data.
For example, a user might create an index named "product_embeddings" on a serving endpoint to enable fast retrieval of similar products based on their vector representations.
Properties
| Name | Meaning |
|---|---|
| Index Name | The name of the vector search index to be created. |
| Endpoint Name | The name of the serving endpoint where the index will be created. |
| Primary Key | The column name that acts as the primary key for the index (default is "id"). |
| Vector Column | The column name that contains the vector data used for similarity search (default "vector"). |
Output
The node outputs JSON data representing the result of the index creation request. This typically includes confirmation details such as the index name, status, and any metadata returned by the API indicating successful creation.
No binary data output is involved in this operation.
Dependencies
- Requires an active connection to the Databricks API with appropriate authentication (an API token).
- The node depends on the Databricks environment being properly configured with a serving endpoint capable of hosting vector search indexes.
- Proper credentials must be set up in n8n to authenticate requests to the Databricks API.
Troubleshooting
Common issues:
- Invalid or missing API token leading to authentication errors.
- Specifying a non-existent or incorrectly named serving endpoint.
- Using an index name that already exists, causing conflicts.
- Providing invalid column names for primary key or vector column.
Error messages and resolutions:
- Authentication failed: Verify that the API token credential is correctly configured and has sufficient permissions.
- Endpoint not found: Check that the serving endpoint name is correct and accessible.
- Index already exists: Choose a unique index name or delete the existing one before creating a new index.
- Invalid column name: Ensure the primary key and vector column names match the schema expected by the serving endpoint.