A single Amazon Redshift cluster is the company’s sole data source and it contains some sensitive fields. A data scientist needs access to some sensitive columns without altering the source data or storing anonymized copies in the database. Which solution requires the least implementation effort?

The correct answer is configuring dynamic data masking policies. Amazon Redshift supports dynamic data masking, which allows you to define policies that mask sensitive data in real time based on user roles or other conditions. This approach directly addresses the requirement to access sensitive columns without altering the source data or storing anonymized copies, as the masking occurs at query time within Redshift itself. This is the least implementation effort because it leverages a built-in Redshift feature. Creating a materialized view requires duplicating data and managing the view, adding complexity. Unloading data to S3 and using Athena or an AWS Glue job involves moving and transforming data outside Redshift, which is more complex and creates additional data copies, violating the requirement of not storing anonymized copies in the database (or elsewhere).