I have attached a workbook, sales-returns-workbook.xlsx, with five sheets. The data dictionary for it is pasted below this message. Please inspect every sheet and tell me: the number of rows in each sheet; the key column of each sheet and whether it is unique; the date range of every date column; and any data-quality issues you find. Then explain which calculations can and cannot be made from this data using the supplied definitions, and why.

Data dictionary:

# Data dictionary - sales and returns workbook (synthetic)

This workbook is fully synthetic. It holds January-June 2026 sales and returns for a made-up retailer.
Dates are written as ISO yyyy-mm-dd.

## Sheets and columns

- Sales: line_id (sales line ID), order_date, product_id (links to Products), region_id (links to Regions), units (whole number, at least 1, units sold on the line). There is no price, revenue or cost column.
- Returns: return_id (unique return event ID), line_id (the Sales line that was returned), return_date, units_returned (whole number, at least 1, never more than the units on that sales line).
- Products: product_id, product_name, category_id (links to Categories).
- Categories: category_id, category_name.
- Regions: region_id, region_name.

## Definitions - apply exactly

1. Deduplicate identical sales-line IDs: exact repeated rows count once.
2. Units are not revenue.
3. Aggregate sales and returns separately BEFORE joining them.
4. Return rate = returned units / sold units.
5. A category with zero sold units has an UNDEFINED return rate (not 0%).
6. Missing months are missing, not zero.

## Notes on the definitions

- Sold units = the sum of units over the deduplicated sales lines.
- Returned units = the sum of units_returned. A return belongs to the product, category and region of the sales line named by its line_id.
- Unless a question says otherwise, work over the whole January-June 2026 period.
