Your applications will be writing their logs to BigQuery for analysis. Each application should have its own table. Any logs older than 45 days should be removed. You want to optimize storage and follow Google-recommended practices. What should you do?

Making the tables time-partitioned and configuring partition expiration at 45 days is the most efficient and Google-recommended approach. This automatically deletes entire partitions (daily, hourly, etc.) when they reach the specified age, optimizing storage and query performance. Configuring the expiration time for the entire table at 45 days would delete the entire table, not just old data, which is not the desired outcome. Relying on BigQuery's default behavior is incorrect as BigQuery does not automatically prune old data based on a custom age. Creating a script is a manual and less efficient solution compared to BigQuery's built-in partition expiration, requiring ongoing maintenance and incurring additional operational overhead.