Overview
This node integrates with an Oracle Database to perform vector similarity searches using stored embeddings. It is designed to work with AI embeddings as input and outputs a vector store object that can be used for further vector-based operations. Common use cases include semantic search, recommendation systems, or any scenario where you want to find similar items based on vector representations stored in an Oracle database table.
For example, you might use this node to:
- Search for documents semantically similar to a query embedding.
- Retrieve nearest neighbors in a recommendation engine.
- Perform clustering or classification tasks by comparing vector distances.
Properties
| Name | Meaning |
|---|---|
| Table Name | The name of the Oracle database table where the vectors are stored and searched. |
| Distance Strategy | The method used to measure similarity between vectors. Options: EUCLIDEAN_DISTANCE, DOT_PRODUCT, COSINE |
Output
The node outputs a JSON object containing a response field. This field holds an instance of a vector store connected to the specified Oracle database table, initialized with the provided embeddings and distance strategy. This vector store can be used downstream in workflows to perform vector similarity queries or other vector operations.
No binary data output is produced by this node.
Dependencies
- Requires an Oracle Database accessible via network with proper credentials (username, password, host, port, service name).
- Needs an API key credential for authenticating to the Oracle database.
- Uses the
oracledbNode.js package to connect and interact with the Oracle database. - Relies on a custom OracleDbVectorStore class that handles vector storage and similarity computations.
- The node expects an AI embedding input connection to receive vector data.
Troubleshooting
- Connection errors: Ensure the Oracle database credentials and network details (host, port, service name) are correct and the database is reachable.
- Authentication failures: Verify that the provided API key credential has sufficient permissions and is valid.
- Table not found: Confirm that the specified table name exists in the Oracle database and contains the expected vector data schema.
- Invalid distance strategy: Use one of the supported options (
EUCLIDEAN_DISTANCE,DOT_PRODUCT,COSINE) exactly as spelled. - Embedding input missing: The node requires an embedding input; ensure it is connected properly.