You have a Snowflake table 'customer data' containing personally identifiable information (PII) such as customer names and addresses. You need to create a de-identified version of this table for analytical purposes, replacing the names and addresses with synthetic data, while preserving referential integrity with other tables that use the original customer IDs. Which combination of Snowflake features and techniques would you use to achieve this securely and efficiently?
Correct Answer: B,D
Options B and D are the best choices. Option B directly addresses the need for a de-identified table by generating synthetic data and using HASH to preserve referential integrity. Secure data sharing provides controlled access. Option D uses external functions to leverage specialized de-identification services. Options A and E are useful for dynamic masking, but do not create a separate de-identified dataset and may still expose PII to authorized users. Option C involves exporting data, which increases the risk of data leakage and is less efficient.