Question 10

Explanation

Option C is CORRECT. Column-level security with policy tags in BigQuery allows for precise control over who can access specific columns containing sensitive information. By creating a taxonomy with policy tags for different sensitivity levels (e.g., "PHI" for Protected Health Information), associating those tags with specific columns in BigQuery tables, and then granting access to tags via IAM, the organization can ensure that only authorized personnel can view patient identifiable information while allowing broader access to non-sensitive columns in the same tables.

Option A is INCORRECT. Creating separate tables introduces data management challenges, potential data duplication, and consistency issues. This approach makes it difficult to perform queries that need both sensitive and non-sensitive data, requiring joins that add complexity and potentially reduce performance.

Option B is INCORRECT. Dataset-level IAM permissions are too coarse-grained, as they control access to entire datasets rather than specific columns. This approach would either grant access to all patient data or none, without the ability to selectively restrict access to just the identifiable information columns.

Option D is INCORRECT. Authorized views with row-level security filter rows based on conditions, not columns. While useful for restricting access to certain patients' records, this doesn't address the need to restrict access to specific columns (like identifiable information) across all patient records.