Google Vertex Vector Search icon

Google Vertex Vector Search

Interact with Google Vertex AI Vector Search

Actions9

Data Point → Upsert

AI-generated

Summary

Insert or update multiple data points in a specified Google Vertex AI Vector Search index by providing their IDs, feature vectors, and optional metadata.

Inputs

  • Location (required) — The geographic location where the Vector Search resources (index) are hosted.
  • Index ID (required) — The identifier of the Vector Search index where data points will be upserted.
  • Data Points (required) — A JSON string representing an array of data points to upsert, each including 'id' and 'featureVector' properties, and optionally 'restricts' and 'crowdingTag'.

Output shape

a single JSON object indicating success status and a message confirming that the data points were upserted successfully.

The operation returns a single object with 'success' set to true and a descriptive message. Errors are handled per item, optionally allowing continuation on failure.

Examples

Example 1: Upsert multiple data points to a Vector Search index with their embedding vectors and metadata.

Specify 'Location' as your data center (e.g., 'us-central1'), provide the target 'Index ID', and supply a JSON array under 'Data Points' with each data point's 'id' and 'featureVector'.

Discussion