Oracle Database Vector Store
AI-generatedSummary
Connects an Oracle database table to an n8n AI vector store, creating the table if it is missing. Use it with an Embedding input to support vector search over documents stored in the configured table.
Inputs
- Embedding (required) — Connect an embedding model used to create vectors for documents and search queries.
- Table Name (required) — Oracle table to use. If absent, the node attempts to create it with ID, text, metadata, and embedding columns; the embedding column dimension comes from the connected model.
- Distance Strategy — Choose EUCLIDEAN_DISTANCE, DOT_PRODUCT, or COSINE to rank vector similarity searches. If unset, the implementation defaults to EUCLIDEAN_DISTANCE.
- Oracle Database credentials (required) — Provide the Oracle connection details through the required oracleDatabaseApi credentials.
Output shape
An AI Vector Store connection for downstream nodes; it does not return ordinary workflow items directly.
Searches embed the query and return up to the requested number of matching documents, each with page content and metadata. The implementation orders results by the selected distance strategy. The node's configured metadata filter is passed to the vector store, but this implementation's similarity search does not apply it.