Convert Parquet to CSV
Convert Parquet files to CSV locally in your browser. Free, no upload, no file size limits — your data never leaves your device.
Loading tool…
Why convert Parquet to CSV?
Parquet is the standard storage format in data engineering — columnar, compressed and fast to scan. But the moment a file needs to leave the data platform, CSV wins: Excel, Google Sheets, CRMs and most business tools still speak CSV first. Converting a Parquet extract to CSV is often the last step before handing data to a non-engineering stakeholder.
How this converter works
- Drop a
.parquetfile into the box above. - The file is registered with an in-browser DuckDB engine by reference — it is not copied, not uploaded, and only the pages needed are read.
- DuckDB streams the data out as CSV with a header row, and the result downloads automatically.
The whole pipeline runs in a Web Worker, so the page stays responsive even for large files.
Things to know before converting
- Types are flattened. Timestamps become ISO-8601 strings, decimals become plain numbers. CSV has no type system, so a round-trip back to Parquet needs type inference.
- Compression disappears. A 100 MB Snappy-compressed Parquet file can easily become a 500 MB+ CSV — columnar compression is that effective.
- Column order is preserved exactly as stored in the Parquet schema.
If you only need to inspect the file rather than convert it, the Parquet Viewer shows schema and data instantly, and the SQL Workbench lets you filter before exporting.
Frequently asked questions
- Is there a file size limit?
- No hard limit is imposed by this tool. Because the file is read by reference inside your browser rather than uploaded, multi-gigabyte files work — the practical limit is your machine's memory for the converted output.
- Is my data uploaded to a server?
- No. The conversion runs inside your browser using DuckDB compiled to WebAssembly. You can disconnect from the internet after the page loads and the conversion still works.
- How are nested columns (lists, structs) converted to CSV?
- CSV is a flat format, so nested Parquet values are serialized to their text representation in a single cell. If you need to preserve structure, convert to JSON or JSONL instead.
- Which Parquet compression codecs are supported?
- Snappy, Gzip, Zstandard, LZ4 and uncompressed files are all supported — the codec is detected automatically from the file metadata.