BigQuery introduces VECTOR_INDEX.STATISTICS to track data drift and ALTER VECTOR INDEX REBUILD for index maintenance, both in Preview. Access Transparency now covers BigQuery data prep in GA. New Preview options for CREATE EXTERNAL TABLE and LOAD DATA include null_markers for NULL strings and source_column_match for column mapping. MATCH_RECOGNIZE clause adds pattern recognition in SQL queries, enhancing analytical capabilities.
Create professional Google Ads performance reports in Google Slides with AI insights and visualizations. Transform Google Ads data into polished presentations featuring AI-generated insights and custom visuals to impress stakeholders and streamline reporting. Use cases include reducing manual reporting time, presenting insights without design skills, maintaining reporting schedules, identifying trends, and sharing campaign insights for strategic decisions.
BigQuery now supports the DISTINCT pipe operator in pipe syntax queries, which returns distinct rows from an input table while preserving table aliases. It works like SELECT DISTINCT or UNION DISTINCT but can be used anywhere in the query. It computes distinct rows based on all visible columns, ignoring pseudo-columns, and does not expand value table fields. This feature is generally available.
Google's Looker Studio added a report abuse feature letting viewers flag inappropriate content, which is then reviewed and removed or hidden. Performance improvements were made for BigQuery data sources, including a short query optimized mode that speeds up queries under certain credential conditions, enhancing response times without changing workflows for data source owners.
You can commercialize BigQuery sharing listings on Google Cloud Marketplace using Analytics Hub. Publishers list data products with pricing and duration options, while subscribers find and use datasets. Required IAM roles and enabling the Analytics Hub API are needed. Listings link Sharing and Cloud Marketplace, creating linked datasets. Subscriptions are managed via Cloud Marketplace orders with standard pricing and revenue sharing.
Google launched an automated data insights feature for BigQuery with Gemini integration that creates natural language questions and SQL queries from metadata, helping analysts explore datasets efficiently. It reveals patterns, checks data quality, and aids analysis without starting from scratch. A Preview feature auto-generates table and column descriptions to improve documentation and discoverability.
Google BigQuery now allows admins to customize the BigQuery Studio interface for projects or organizations. It supports Spanner external datasets with enhanced security features. Data loading options include time zone and date/time formatting parameters in preview. A new Apache Spark demo notebook is available to demonstrate serverless Spark capabilities. These updates improve integration, flexibility, and user experience.
Looker Studio enables Code Interpreter by default for Pro users when Gemini in Looker and Trusted Tester settings are active in the Google Cloud project. It converts natural language questions into Python code for advanced analysis and visualizations. Admins can control the feature via User Settings. Requirements include a Pro subscription, Google Cloud project, and enabled settings, enhancing AI-powered analytics in Looker Studio.
Google BigQuery offers continuous queries for real-time data processing and ML inference, outputting to BigQuery, Pub/Sub, or Bigtable. It supports multiple ingestion methods, monitoring, and autoscaling. Use cases include personalized messaging, anomaly detection, event-driven pipelines, data enrichment, and reverse ETL. Spanner supports cross-regional federated queries from BigQuery, enabling cross-region queries without egress fees in preview.
BigQuery now allows scheduling automated data transfers from Snowflake using the Data Transfer Service, currently in preview, staging data in the same cloud or Amazon S3 for AWS accounts. It supports cross-region batch loading and exporting data across any regions with simple commands, now generally available. Vector indexes support the TreeAH type with Google's ScaNN algorithm for efficient batch processing of many embeddings, also generally available.
