Sample Excel File

This sample is a minimal Office Open XML workbook: one worksheet named employees, a header row, and 30 data rows — no formulas, no merged cells, no formatting beyond defaults. It was written with openpyxl, the same library pandas.DataFrame.to_excel() uses, so it represents the .xlsx files that data pipelines actually emit.

Under the hood, .xlsx is a ZIP archive of XML parts. The workbook manifest points at xl/worksheets/sheet1.xml, which stores cells by coordinate (A1, B2, …); repeated strings like department names live once in a shared-strings table and are referenced by index. That indirection is why parsing Excel is an order of magnitude more involved than parsing CSV — and why libraries such as openpyxl, SheetJS, and calamine exist.

Two type behaviors worth testing against this file:

• salary is a true numeric cell. Spreadsheet apps will right-align it and let you aggregate immediately — no text-to-number coercion step.
• start_date is stored as an ISO 8601 string, not an Excel serial date. That is deliberate: serial dates (days since 1900, with the famous fictitious leap-day bug) are a classic source of off-by-one and format-guessing errors when files cross tools. A string column round-trips losslessly through every converter.

The single-sheet, header-plus-rows shape is the format every importer expects, which makes this file a good smoke test for upload flows, read_excel() wrappers, and the site’s own Excel to CSV converter. Going the other direction, CSV to Excel produces a workbook with the same structure from any CSV.

All eleven sample formats in this section express the same 30 employee rows — convert this workbook to CSV and it should match the CSV sample byte-for-byte on values.