Data Quality: Persist and Inspect Corrupt Rows in Shortcut Transformations

Shortcut Transformations today silently drops corrupt rows during file-to-Delta ingestion - users see a count of rejected rows in telemetry but have no way to inspect what failed, which file it came from, or when it was rejected. It's estimated that 35% of all jobs encounter corrupt rows that are silently discarded. This feature persists corrupt rows alongside valid data in the destination Delta table with source file path and ingestion timestamp, enabling direct inspection via SQL/Power BI, downstream remediation in silver-layer pipelines, and compliance audit trails. Opt-in via a simple feature flag with zero regression for existing shortcut transformations whereas it would be supported for all new shortcut transformations by default. The feature is applicable to CSV, JSON, and Excel files