Parquet to Excel converter — how it works
Converting Parquet (a columnar, compressed analytical file format) to Excel (an Excel workbook) happens entirely in your browser. Because Parquet is a binary columnar format, a text editor or spreadsheet cannot open it directly — it needs a reader that understands the schema. Excel keeps the data openable by non-technical colleagues in a spreadsheet, with the usual worksheet row ceiling of 1,048,576 rows per sheet.
- Cost
- Free; no account required
- Processing
- Local, inside the browser tab
- Privacy
- File contents are not uploaded
Steps
-
Open the Parquet file
Drop or select a .parquet file. Excel cannot read this format directly, which is the reason this page exists.
-
Cut it down to something a sheet can hold
A worksheet stops at 1,048,576 rows, and an analytical Parquet file often carries more. Filter, or aggregate with SQL, until the result fits — check the row count in the status bar before exporting.
-
Export as XLSX
Open the Export panel, choose Excel, and download the .xlsx. It opens in Excel, LibreOffice, Numbers or Google Sheets.
Details that matter
The same thing in the DuckDB CLI
INSTALL excel;
LOAD excel;
COPY (SELECT * FROM read_parquet('data.parquet')) TO 'data.xlsx' (FORMAT xlsx, HEADER true);The CLI route needs DuckDB’s excel extension installed once, and it writes .xlsx only — the legacy .xls format is not supported there. Worth knowing if you are automating; otherwise this page saves the setup.
Checking that nothing is uploaded
Nothing here needs trusting blind. Open your browser’s developer tools at the Network panel, then convert a file: the requests you see are this page’s own scripts and styles and the DuckDB engine. Your workbook data is not among them.
When a spreadsheet is the wrong destination
Excel is for people to read, not for pipelines to consume. Sending it a large analytical extract means losing the schema, hitting the row ceiling, and inviting silent edits nobody can audit. If the recipient is a machine, or the data will be re-imported later, keep it as Parquet and send a filtered CSV only if you must.
Limits and behaviour worth knowing
One worksheet holds 1,048,576 rows and 16,384 columns. Numbers, dates and booleans are written as native Excel cell types, so sorting and formulas behave. Nested struct and list columns are serialised to text, spreadsheets being flat. Excel stores numbers as 64-bit floats, so identifiers longer than 15 digits lose precision on display — export those as text if they must stay exact.
Frequently asked questions
Can Excel open a .parquet file directly?
No. Excel has no Parquet reader, which is why the usual advice is a Power Query connector or a Python script. Converting here is the shorter path: drop the file, export XLSX, open that.
What if the Parquet file has more rows than Excel allows?
A sheet caps at 1,048,576 rows. Filter or aggregate with DuckDB SQL first until the result fits, or export CSV instead — CSV has no such ceiling, though Excel will still refuse to display beyond it.
Do Parquet dates arrive as real Excel dates?
Yes. Date and timestamp columns are written as native date cells rather than text, so they sort chronologically and work in formulas without re-parsing.
Can I export several sheets at once?
Each open tab can be exported. Split the data into tabs first — by running a query per slice, for example — and export them as separate sheets.