Sample Data Files for Download — CSV, JSON, XML, SQL, SQLite, YAML, GeoJSON
17 formats · 32 files · direct links · no ads · free for any use
The structured files share one fictional dataset — orders from a small Canadian distributor with customers, regions, products, quantities and totals — so the same records appear as CSV, JSON, XML, SQL inserts and a SQLite database. That lets you test an import in one format and verify it against another.
Edge cases are included on purpose: a CSV with a UTF-8 BOM, one with no header row, one with quoted fields containing commas and line breaks, a semicolon-delimited European variant, and JSON Lines for streaming parsers. Row counts run from 10 to 100,000 so you can profile performance as well as correctness.
Choose a format
Sample CSV files
Comma-Separated Values
Sample TSV files
Tab-Separated Values
Sample JSON files
JavaScript Object Notation (and JSON Lines)
Sample XML files
Extensible Markup Language
Sample YAML files
YAML Ain't Markup Language
Sample TOML files
Tom's Obvious Minimal Language
Sample SQL files
SQL script (schema + data)
Sample SQLite files
SQLite database file
Sample GeoJSON files
GeoJSON (RFC 7946)
Sample LOG files
Web-server access log (nginx/Apache combined format)
Sample vCard files
vCard contact file (.vcf)
Sample iCalendar files
iCalendar (.ics) calendar file
Sample SRT files
SubRip subtitle file
Sample WebVTT files
Web Video Text Tracks (.vtt)
Sample STL files
Stereolithography 3D model (ASCII)
Sample OBJ files
Wavefront OBJ 3D model
Sample Binary files
Raw binary test file (random bytes)
Frequently asked questions
Is the data real?
No. Customers, products and orders are fictional and generated from a fixed random seed. Names are invented, addresses are generic Toronto street addresses, and no value corresponds to a real company or person, so the files are safe to commit to a public repository.
Can I use these as fixtures in automated tests?
Yes — that is what they are for. Every file is deterministic and its SHA-256 is published on the page, so you can pin the hash in your test and detect if the fixture ever changes. Direct links have open CORS headers and long cache lifetimes.