OIMRQS Ops Sample Artifact: Data Cleanup Handoff

This is a public sample of the kind of handoff a focused cleanup pass should leave behind. It does not use client data.

INPUT

- File: orders-export.csv
- Rows received: 2,480
- Columns received: order_id, customer_name, email, total, created_at, status
- Known issue: duplicate order IDs, mixed date formats, blank emails, totals stored as text.

WORK DONE

1. Normalized dates into YYYY-MM-DD.
2. Trimmed whitespace from names and emails.
3. Converted totals into numeric decimal values.
4. Flagged invalid emails instead of deleting them.
5. Split duplicate order IDs into a rejected rows file for review.

OUTPUT

- Clean file: orders-clean.csv
- Rejected rows: orders-rejected.csv
- Summary: cleanup-summary.json

VALIDATION COMMAND

python3 cleanup_orders.py orders-export.csv --out orders-clean.csv --rejects orders-rejected.csv

CHECKS

- Input rows: 2,480
- Clean rows: 2,431
- Rejected rows: 49
- Duplicate order IDs found: 31
- Invalid emails flagged: 18
- Totals parsed successfully: 2,480

HANDOFF NOTES

- No rows were silently deleted.
- Rejected rows are separated so the client can review edge cases.
- The script can be rerun on the next export with the same command.
- If the source system can export stable date and numeric formats, the cleanup step gets simpler.