A data warehouse must ingest partner data that contains PII but store only non-PII. How to process and store without retaining PII?

This option directly addresses the requirement to ingest PII, remove it, and store only non-PII. Dataflow is excellent for streaming or batch data processing, and integrating Cloud DLP within the pipeline allows for real-time PII detection and redaction before the data ever reaches persistent storage in BigQuery. This ensures PII is never retained in the data warehouse. The option to store PII in Cloud Storage with a retention policy still retains PII, which is explicitly against the requirement. Requiring partners to upload to Cloud Storage with Bucket Lock and then processing is less efficient than direct ingestion and processing, and Bucket Lock doesn't prevent PII from being stored initially. Asking partners to import directly into BigQuery and then skipping columns is risky, as PII would temporarily reside in BigQuery and column-skipping might miss PII in other fields.