Question

Your team is migrating a 40 TB transactional dataset from an on-premises Oracle database into BigQuery using a one-time bulk load followed by change-data-capture streaming. After the initial bulk load completes, the compliance team requires proof that every source row was faithfully transferred with no truncation or silent data corruption before the cutover is approved. Which validation approach best establishes end-to-end data fidelity?

Accepted Answer

Compare source and destination table row counts, and additionally compute and compare aggregate checksums (e.g., hash totals of key columns) on both systems for each partition.

Answer

Enable Cloud Monitoring dashboards on the Dataflow migration pipeline and confirm there are no error logs during the bulk load window.

Answer

Run a SELECT COUNT(*) on the BigQuery tables and confirm the number is greater than zero for every migrated table.

Answer

Sample 100 random rows from BigQuery and manually verify their values against the Oracle source.