code-check-data-integrity

Checking database migrations for NOT NULL backfill gaps involves scanning schema diffs to detect new columns added without defaults that would fail on existing rows. The review flags missing backfill steps and ranks findings by severity to prevent production data breakage.

Preventing orphaned child rows during database deletes requires auditing destroy paths for unhandled cascade behavior. The review flags destructive removals that leave orphaned children or external side effects, providing actionable recommendations for safe cascade behavior and rollback sequencing.

Finding code that assumes uniqueness without database schema enforcement involves reviewing persistence layers for unenforced constraints. The audit detects mismatches between application-level uniqueness assumptions and the actual database schema, ranking these vulnerabilities by severity.

Detecting seed or fixture drift after a schema update requires comparing fixture files against the modified schema to catch missing or removed required fields. This validation highlights fixture drift to ensure test data matches current persistence requirements.

Scanning for destructive and irreversible migration operations like DROP COLUMN, DROP TABLE, or TRUNCATE is necessary during code review or pre-deploy validation. This highlights destructive migration irreversibility to prompt backups and safe rollback sequencing.

Yes, this data integrity check works with background jobs and imports by auditing whether persisted state changes preserve existing rows, constraints, and related child data. It applies during code review when diffs indicate mutating or destructive persistence changes.