You need a cost-effective analytics environment so data scientists can query ORC files with Hive-style partitioning using SQL, without affecting on‑prem Hive. What should you do?

The correct approach is to copy the ORC files to Cloud Storage and create external BigQuery tables. This is cost-effective because BigQuery external tables query data directly from Cloud Storage without needing to import it, and BigQuery is serverless, eliminating cluster management. It supports Hive-style partitioning and SQL, meeting all requirements without impacting on-prem Hive. Importing to Bigtable is incorrect because Bigtable is a NoSQL wide-column store, not suitable for SQL analytics on ORC files. Importing to BigQuery tables would incur storage costs for duplicated data and requires an import process, which is less cost-effective than external tables. Deploying a Dataproc cluster is an option for querying ORC files, but it involves managing a cluster, which is less cost-effective and serverless than BigQuery for ad-hoc SQL queries.