Sample Parquet Generator

110,000 rows

Each row is fictional fixture data over a fixed, mixed-type schema.

The file is built and downloaded entirely in your browser with the Web Crypto API — nothing is uploaded. It is a real, valid Apache Parquet file you can open in the Parquet Viewer or load into any data tool.

Schema

ColumnTypeDescription
idINT64Sequential row identifier
nameSTRINGFictional full name
emailSTRINGFictional email address
activeBOOLEANRandom true / false flag
scoreDOUBLERandom value 0–100
created_atTIMESTAMPRandom timestamp in the past two years

Preview — first 5 rows

idnameemailactivescorecreated_at
1Blair Hayesv7mrfubc9-m.4pw_i0@example.orgfalse11.442026-08-20T06:20:46.925Z
2Rowan Bennettbkqzdn2@example.comtrue31.612024-10-17T02:26:09.986Z
3Jordan Morgandg_7iyyy8@example.comfalse11.372024-11-10T05:57:35.598Z
4Rowan Millerobbsh-1mr9@example.comtrue64.902025-01-10T06:35:50.445Z
5Sage Lawsontao4v60crnaeb6@example.netfalse16.872024-10-09T16:01:58.280Z

Fresh random values are generated on each download, so your file will differ from this preview.

Dummy Parquet files for testing data pipelines

Generate a small, valid Apache Parquet file with a chosen number of rows without hunting for a sample to download. Use it to exercise a loader, viewer, or ETL step, check that a columnar reader handles INT64, BOOLEAN, DOUBLE, and TIMESTAMP columns, or hand a data pipeline a throwaway fixture. Every file carries clearly-fictional data over one fixed schema, so it is always safe to share.

How it works

Pick anywhere from 1 to 10,000 rows and download. The file is encoded to real Parquet bytes in your browser — column-oriented, PLAIN-encoded, and bracketed by the PAR1 magic — with the values drawn from the Web Crypto API. Nothing is uploaded and no server is involved. Open the result in the Parquet Viewer to confirm the schema and row count round-trip.