Convert CSV to Parquet
You have a CSV. The next step in the pipeline wants Parquet. You can do that in the browser without standing up Spark or pandas.
Convert in the tab, not on a server
Parquet Fox opens the CSV locally. SQL runs in DuckDB WebAssembly in this tab. The file is not uploaded to a Parquet Fox server for open, query, or export.
Caps we actually enforce: Parquet files up to 500 MB; CSV and JSON up to 2048 MB compressed on disk. A highly compressible CSV can still expand in RAM. If the tab stalls, filter with SQL before you export.
SQL before export
Run SQL in the tab first if you need typed columns, a subset of rows, or a cleaned header. Then export Parquet. The writer uses the same DuckDB export path as CSV and JSON.
- Open a file on the homepage and drop the CSV. TheCSV viewer landing uses the same app if you want CSV-first copy.
- Optionally run SQL to cast types or drop unused columns.
- Use Export and Parquet.
When this is the wrong convert
Stay on CSV when a colleague only needs Excel. Use a local DuckDB or pandas job when the file is larger than the 2048 MB compressed cap, or when you need a repeatable pipeline.
FAQ
- Can I convert CSV to Parquet without uploading the file?
- Yes. Parquet Fox reads the CSV in your browser tab. The file is not uploaded to a Parquet Fox server for open, query, or export.
- Why convert CSV to Parquet at all?
- Parquet is columnar and compressed. It is smaller to store and faster to query than a wide CSV once you are past a one-off spreadsheet. If you only need a peek, stay on CSV.