A CSV load maps values to the wrong columns because the source layout changed. What should be corrected before retrying?

1. Increase CPU without changing parsing configuration
2. Disable all error reporting
3. The source-to-target mapping and file format assumptions
4. Rename the target table only