# Data QA audit: dirty-orders.csv

- Demonstration data: synthetic
- Rows reviewed: 7
- Columns: 5
- Intended key: `record_id`
- Disposition: blocked pending duplicate-key and missing-amount review

## Issue register

1. **High · duplicate key · `record_id=1003`:** Rows 4 and 5 use the same intended business key but contain different amounts and status values. Confirm whether this is a revision, split event, or erroneous duplicate.
2. **High · exact duplicate:** Rows 7 and 8 are identical. Remove only in a derived clean file after confirming the duplication was not intentional.
3. **High · duplicate key · `record_id=1006`:** The intended key repeats at rows 7 and 8.
4. **Medium · missing value · `amount`:** One of seven rows is blank. Do not silently coerce the value to zero.
5. **Medium · invalid date text · `order_date`:** `not-a-date` does not match the stated ISO date contract.
6. **Low · outer whitespace · `customer`:** ` Silverline Co ` contains leading and trailing whitespace.

## Controlled correction plan

- Preserve this input unchanged.
- Ask the source owner to resolve `record_id=1003` and the missing amount.
- Mark the exact repeated `1006` record for removal in the derived file after approval.
- Trim outer whitespace and normalize the controlled status vocabulary in the derived file.
- Re-run this audit and reconcile record count and amount total before handoff.

## Limits

This audit checks structure and declared rules. It cannot confirm that amounts, customer names, or dates are factually correct, and it does not treat every unusual value as an error.