Google announced a new general availability (GA) release for BigQuery, introducing the ability to monitor stored column usage information specifically for query jobs performing vector searches.
The update introduces the StoredColumnsUsage feature, which provides detailed insights into how stored columns are utilized during vector search operations. This enhancement offers three key components:
- storedColumnsUnusedReasons: Provides explanations when stored columns remain unused during queries
- isQueryAccelerated: Indicates whether stored columns accelerated the query execution
- baseTable: Identifies the primary table referenced in the query
JSON Implementation
The feature implements a JSON structure that tracks these metrics. The system monitors and reports the status of stored columns through a comprehensive object format that includes:
- Detailed reasoning for unused stored columns
- Boolean flags for query acceleration status
- Reference information for the base table being queried
This update represents a significant enhancement to BigQuery's vector search capabilities, providing developers and data engineers with more granular insights into query performance and stored column utilization.