You are developing a real-time prediction engine that processes streaming files. These files, which may contain personally identifiable information (PII), are ingested into Cloud Storage and eventually stored in BigQuery. To protect this sensitive data, you need to ensure that it is masked while still preserving its referential integrity, as names and emails are frequently used as join keys in your database. How can you utilize the Cloud Data Loss Prevention API (DLP API) to guarantee that the PII data is not accessible by unauthorized individuals?