Synthetic datasets for analysis and ETL testing (CSV, JSON, JSONL, Parquet, XLSX, XML, YAML, TSV)

Download clean, deterministic multi-row datasets across three domain schemas—Sales Orders, Employee Directory, and Weather Station Telemetry—at 100 and 1,000 rows, complete with column-by-column data dictionaries and SHA-256 checksums.

Why deterministic synthetic datasets matter for QA

Testing an ETL pipeline, a business-intelligence dashboard, a Pandas/Polars script, or a database bulk loader with production customer data violates privacy regulations, while testing with random Lorem Ipsum strings fails to exercise numeric aggregations, date parsing, boolean filtering, or group-by joins. Every dataset on this page is generated from a fixed mathematical seed using three realistic relational schemas so that your unit tests can assert exact row counts, exact column sums, and reproducible SHA-256 file hashes across CSV, TSV, JSON, JSONL, XML, YAML, Excel XLSX, and Apache Parquet.

Need custom row counts up to 50,000 rows or deliberate edge cases such as embedded commas, line breaks, and missing values? Open the interactive CSV, JSON and TSV Test Data Generator.

Download synthetic dataset files (39 files)

Synthetic dataset downloads
Dataset fileFormat & MIMEExact sizeSHA-256 checksumDownload
employees-1000rows.csv .csv (text/csv) 56.94 KB58,304 bytes 44fb7ce4d84624188107e390360e4d6f60586cc782896af99e7b52f0ded63e1c Download
employees-1000rows.json .json (application/json) 196.52 KB201,236 bytes 1f65c5cccd49a6e12d42cf6ac2cd9340bd407881477993bf0e7b9354959eef21 Download
employees-1000rows.jsonl .jsonl (application/x-ndjson) 162.34 KB166,233 bytes 8d213c012804506d94b28b1d6124de71c030a130b37731a6e0116e8ff8b6dd3f Download
employees-1000rows.parquet .parquet (application/vnd.apache.parquet) 16.02 KB16,402 bytes ebb02c8e1e65dd7d6548940ce7589845cb3c262d33e0fd2c9adbe8ae5a80bd7c Download
employees-1000rows.tsv .tsv (text/tab-separated-values) 56.94 KB58,304 bytes 7ab356ca475f37778a83a78236dd541649b4839b7a49c3684aada5b024b1878f Download
employees-100rows.csv .csv (text/csv) 5.76 KB5,894 bytes 3ddaaae0d9872563cc3dbee554c52e1f70369ca91fc9cc1bec95ecd878b8ad2a Download
employees-100rows.json .json (application/json) 19.65 KB20,126 bytes ec3e08f2361ef44a30c5272490256b559d1a4b84beaffe87c2fcda7ce7ec9031 Download
employees-100rows.jsonl .jsonl (application/x-ndjson) 16.23 KB16,623 bytes 4a35b459dcd832bce3e108fa63b4638118425ef035e1b9e64a04db17d3f73c4e Download
employees-100rows.parquet .parquet (application/vnd.apache.parquet) 5.71 KB5,850 bytes 5b1565386839b134eeadd80a2ba7b3d7b42bf74fa9c6dc822050f85d183e1c2c Download
employees-100rows.tsv .tsv (text/tab-separated-values) 5.76 KB5,894 bytes 3cd7fb4f29f3c72f7b7ee7c65d66367732979b651449dd853db549b3b8fc0ac6 Download
employees-100rows.xlsx .xlsx (application/vnd.openxmlformats-officedocument.spreadsheetml.sheet) 8.74 KB8,951 bytes 777af668f45cd893a8307eb4b485bbca4fc704010c295ea183116de2d04024bc Download
employees-100rows.xml .xml (application/xml) 26.65 KB27,287 bytes 2b8a80dfba8de1ce77f2b28f88ac217c5482a4801883ddbe97420d28f7d64b61 Download
employees-100rows.yaml .yaml (application/yaml) 16.87 KB17,272 bytes fcb447dbd91a17b842d71992dae11fdb0dcc702c35be3f7ab03c53a5af05fddb Download
sales-1000rows.csv .csv (text/csv) 65.87 KB67,453 bytes 15857c1499c3d4416670a312d64f06b4b6794e0d7e2176c1a9c7cd0069441a67 Download
sales-1000rows.json .json (application/json) 220.10 KB225,378 bytes 119424225c87c809746aa8e54eef6372398c6e54937ab8fa3f8d26a785c32a34 Download
sales-1000rows.jsonl .jsonl (application/x-ndjson) 182.01 KB186,375 bytes 84f7f10ae49a1ea15a7ac8be5020cd670b9a0a8873bb2284c916f72420a2eec0 Download
sales-1000rows.parquet .parquet (application/vnd.apache.parquet) 12.44 KB12,737 bytes c6a066ae02f5f4168c4d93d369ed0cde7660b9e567ead214a942922f02388005 Download
sales-1000rows.tsv .tsv (text/tab-separated-values) 65.87 KB67,453 bytes 885a0deb977469d6c5644692ed088c3e52b327dd8ea619964ad7e02106382f1f Download
sales-100rows.csv .csv (text/csv) 6.65 KB6,811 bytes 5c41b5e3e2ee03369b86b653ca6d12895d48b202dad5ba8ec51bbcb12e3c3065 Download
sales-100rows.json .json (application/json) 22.01 KB22,536 bytes c8ebd5beb7738d70a06b4fcc9862ca6dd5ecc93b5814f4bd78d52fb179eb536d Download
sales-100rows.jsonl .jsonl (application/x-ndjson) 18.20 KB18,633 bytes a4638bdfbdc3de0fad51c131863c31e6b7f5b0a27033215c329747c2d2e2beab Download
sales-100rows.parquet .parquet (application/vnd.apache.parquet) 6.67 KB6,832 bytes 2642d2aacff42af41d2890d13a4272a9155bcc37d030dc89bfb2dc14041f1229 Download
sales-100rows.tsv .tsv (text/tab-separated-values) 6.65 KB6,811 bytes 70c76eb32da63542cd26f4c71e53775a20f96715d5d51426be6095d0c9b69e14 Download
sales-100rows.xlsx .xlsx (application/vnd.openxmlformats-officedocument.spreadsheetml.sheet) 9.36 KB9,584 bytes de68cbbe103af5600f1996aad7c76398e1f8f9d5c5bfaadc0b93b83877cb7f06 Download
sales-100rows.xml .xml (application/xml) 29.58 KB30,289 bytes fd4a93e169c7bae16b25d44bdec2276753efb9f61c1265049ae6d410cf2d1e69 Download
sales-100rows.yaml .yaml (application/yaml) 18.92 KB19,374 bytes fc16f126a6d9a7c1caa207fdcbeccdc41154eee527fdef262ed3f03161c0808d Download
weather-1000rows.csv .csv (text/csv) 56.44 KB57,795 bytes 5e6096fe54b4a162acaa732688281130ca2010e87e4b24f5160db5a16615fa81 Download
weather-1000rows.json .json (application/json) 198.95 KB203,722 bytes 36f8802dec6219f2c90d5dee71523f3b437c5480d8debdf13d1e2a2ebcc85b3a Download
weather-1000rows.jsonl .jsonl (application/x-ndjson) 164.76 KB168,719 bytes 5a133db9cdb0837b10ce5dab3735134081500841ef706af25b5cf67e3d23e95a Download
weather-1000rows.parquet .parquet (application/vnd.apache.parquet) 7.48 KB7,660 bytes 351151491fcb85bcb57a1949f946386d41d5be6ae2bf155c4d860784d9f02de6 Download
weather-1000rows.tsv .tsv (text/tab-separated-values) 56.44 KB57,795 bytes 5f4cfa5ba820fc559d8da3737ce1b7f72f69f667b6f4f79c6bd899ff2d25949c Download
weather-100rows.csv .csv (text/csv) 5.71 KB5,850 bytes f0d7bbad7287f338cac31cc5b2701119a8048c8423de4a22a4bd22ebc9e9ef27 Download
weather-100rows.json .json (application/json) 19.90 KB20,377 bytes 0f4462367822ffed675b4a6c5d7d3bba89e1024a1b982703565f744187823d85 Download
weather-100rows.jsonl .jsonl (application/x-ndjson) 16.48 KB16,874 bytes 757e27bee9dc0f63aa8b0aa944b73584ecb33c0e15c5cec8e06b9348c0194f3d Download
weather-100rows.parquet .parquet (application/vnd.apache.parquet) 6.70 KB6,863 bytes a1d90fc3e170e42adc2798373926b6b635763c68f92aeab9bee8bc1416859630 Download
weather-100rows.tsv .tsv (text/tab-separated-values) 5.71 KB5,850 bytes 7580056c44b6e4d909a30e31c897f85bde7d72ac66c8f9d957a5be8f531226f5 Download
weather-100rows.xlsx .xlsx (application/vnd.openxmlformats-officedocument.spreadsheetml.sheet) 9.21 KB9,430 bytes 0677b4fc476607eff442abc18ab90fe4adca6edc6de6b48348784a44c93123a7 Download
weather-100rows.xml .xml (application/xml) 27.57 KB28,234 bytes 78fd00360d8bf7bbff12e0972fb5df734d1923f84e3800ef956995e3f25e42ba Download
weather-100rows.yaml .yaml (application/yaml) 17.11 KB17,519 bytes 6964a2e243d6e2770bbca21345ff2ef7f1e7ab95ec3349049c6bd6772eb59d6d Download

Data dictionaries for all three dataset schemas

1. Sales Orders schema (sales-100rows.* and sales-1000rows.*)

Models an e-commerce and B2B order ledger suitable for testing currency rounding, regional group-by aggregations, and boolean payment filtering.

Column nameData typeExample valueValidation rules & domain range
order_idString (UTF-8)ORD-2026-00001Sequential primary key formatted as ORD-2026-NNNNN
order_dateISO 8601 Date2026-01-01Valid calendar dates cycling months 01–12 and days 01–28
regionCategorical StringNorth AmericaOne of: North America, Europe, Asia-Pacific, Latin America
categoryCategorical StringHardwareOne of: Hardware, Software, Cloud Services, Accessories
quantityInteger (INT64)2Whole integer between 1 and 8 inclusive
unit_price_usdFloat / Decimal17.50Unit price in USD between 15.00 and 112.50 in 2.50 increments
total_usdFloat / Decimal35.00Exact product quantity * unit_price_usd for cross-column assertion checks
settledBooleantruetrue for 80% of rows; false on every 5th row

2. Employee Directory schema (employees-100rows.* and employees-1000rows.*)

Models an organizational HR directory using strictly synthetic identifiers and levels—zero real people or personal data.

Column nameData typeExample valueValidation rules & domain range
emp_idString (UTF-8)EMP-1001Unique synthetic identifier starting at EMP-1001
full_nameString (UTF-8)AlexUser0001 TaylorSynthetic placeholder name guaranteed not to match real individuals
departmentCategorical StringQuality AssuranceOne of: Engineering, Quality Assurance, Operations, Product, Finance
levelStringL3Job band from L2 through L6
remoteBooleantrueAlternating boolean flag for remote vs. on-site filtering tests
tenure_monthsInteger8Integer tenure between 1 and 84 months
annual_Band_usdInteger77000Synthetic compensation band integer between 65000 and 125000

3. Weather Station Telemetry schema (weather-100rows.* and weather-1000rows.*)

Models hourly IoT sensor readings for time-series ingestion, floating-point averaging, and timestamp parsing tests.

Column nameData typeExample valueValidation rules & domain range
station_idStringWX-00220 station identifiers from WX-001 to WX-020
observed_utcISO 8601 UTC Timestamp2026-10-01T01:00:00ZUTC timestamp string ending with Z
temp_cFloat9.0Ambient temperature in degrees Celsius (6.1 to 30.1 °C)
humidity_pctInteger36Relative humidity percentage (35% to 94%)
pressure_hpaFloat1005.1Barometric pressure in hectopascals (1004.0 to 1027.1 hPa)
wind_kphFloat5.3Wind speed in kilometers per hour (4.0 to 48.2 km/h)
conditionCategorical StringPartly CloudyOne of: Clear, Partly Cloudy, Overcast, Light Rain

Compare the parsing performance and schema preservation of these formats in our guide: Dataset formats compared: CSV, TSV, JSON, JSONL, Parquet, Avro, XML and YAML.

Frequently asked questions

Can I verify that my parser produces identical DataFrames from the CSV, JSON, XLSX, and Parquet versions?
Yes. For a given domain and row count (such as sales-100rows), the CSV, TSV, JSON, JSONL, XML, YAML, XLSX, and Parquet files contain the exact same records and field values.
Do the employee or sales datasets contain any real people or companies?
No. All records are mathematically generated using placeholder identifiers (such as AlexUser0001 Taylor and ORD-2026-00001).