How to convert Parquet to CSV
Load a local .parquet file, choose the columns and formatting rules, then select Download CSV. Parquet to CSV Converter reads the file's footer, schema, and row groups directly. Download type report (sidecar) adds the source metadata that CSV cells cannot carry. An unreadable footer or unsupported row-group feature or encoding stops processing with an error.
How to use Parquet to CSV Converter
- Drop a
.parquetfile into Drop a .parquet file here or select browse files. - Check the columns to include, then use the arrows to set their order. Select at least one column before exporting.
- Choose a nested column policy and set the CSV output controls.
- Review Preview, then select Download CSV. Select Download type report (sidecar) to save the JSON metadata file as well.
How Parquet to CSV handles nested columns
The column list shows top-level fields and marks nested fields. For a group, struct, list, or map, choose JSON-stringify whole value, Flatten selected paths (dotted, comma-separated), or Omit (report count of omitted values). JSON-stringify whole value puts the nested value in one CSV cell.
Parquet value: null
Null representation: (empty)
CSV field: (empty)Flatten selected paths creates one output column for each path you enter. A path with no value uses Null representation. Omit writes an empty cell for the nested value and records its count in the sidecar report.
Parquet CSV delimiter, dates, and binary settings
Under Delimiter, choose Comma (,), Tab, Semicolon (;), Pipe (|), or Custom. A custom delimiter and Quote character each accept one character. Quote all fields and Include header row change the CSV layout. Encoding has UTF-8, UTF-8 with BOM (Excel-friendly), and UTF-16LE; Newline has CRLF or LF.
Null representation supplies text for missing values. Date/time format has ISO 8601, Date only (YYYY-MM-DD), Unix epoch (ms), Unix epoch (s), and Locale string. Decimal formatting has As decoded, Fixed to column scale, and Plain number. Binary column handling can write Hex string (0x...), Base64, an empty field with Omit (report count), or stop the export with Error (stop export).
Preview, export, and type report
Preview shows up to the first 25 rows with the current columns and settings. The export reads each row group, reports its progress, assembles the CSV, and enables Download CSV when the file is ready. Error (stop export) stops at a binary value that reaches that policy.
Download type report (sidecar) saves JSON with the source file name, row count, row-group count, source column metadata, settings, selected and exported columns, and omission counts. Keep it beside the CSV when the source types or export choices need to be checked later.
Where Parquet to CSV conversion comes up
- Select top-level columns, choose Include header row, and open the CSV in a spreadsheet.
- Enter
address.cityandaddress.zipas dotted paths to create separate flattened columns. - Keep the sidecar report with the CSV to retain the source Parquet types and export settings.
When to use Database Convert Tools
Use this page when the source is Parquet and the target is CSV. It belongs to Database Convert Tools. The selected file is read locally in the browser, and this page does not upload or modify it.
Frequently asked questions
Can I convert a Parquet file to CSV without converting it to JSON first?
Yes. Parquet to CSV Converter reads the footer, schema, and row groups directly, then writes CSV rows from decoded values.
How do I export only selected Parquet columns?
Check the columns you want in the column list and move them with the arrows. The CSV uses that selected order.
How are nested Parquet values written to CSV?
They are JSON-stringified by default. You can flatten entered dotted paths or omit the nested value and record its count in the sidecar report.
Does Parquet to CSV Converter upload my file?
The selected file is read locally in your browser. The CSV and optional report are generated there for download, and the source file is not modified.