Open a Parquet file and edit it in your browser — change cells, add or delete rows, rename or drop columns — then download a new Parquet. Files never upload.
How the Parquet editor works
Open a .parquet file and edit it like a small spreadsheet: change cells, add or delete rows, rename or drop columns, then download a new .parquet — the original is never touched. On write, column types are re-inferred from the edited values and the file is written fresh with your chosen codec. Everything runs locally in your browser.
What to know before editing
Types re-infer on write — text in a numeric column turns it STRING.
Files load up to the first 100,000 rows; the download contains only the loaded rows.
Nested structs and lists are not editable — this tool is for flat tables.
Output is standard Parquet, readable everywhere.
Bulk edits are a query away
For mass updates, computed columns, or filtering thousands of rows, use SQL instead: open the file in the query editor, transform with SELECT, and export the result as Parquet.