Open Parquet Without Python

You have a Parquet file and no Python

Someone sent you a .parquet file. You do not have Python. You cannot pip install pandas. The laptop may be locked down, or you only need a five-minute peek before a meeting.

That constraint is common. It does not mean the file is unreadable.

You do not need Python

Python is not required to open a Parquet file. Local options that skip a Python environment:

• A browser-local viewer that runs SQL in the tab (no install)
• The DuckDB CLI binary, if you can install one program
• A desktop Parquet viewer, if you can install a GUI app

Pick the path that matches what you are allowed to install. The rest of this page starts with the no-install case, then the two short install-a-binary cases. For the full method table—including pandas—see How to Read a Parquet File.

Open Parquet in the browser (no Python, no install)

If you cannot install Python or any other tool, use a viewer that processes the file in your browser tab. The query engine runs on your device. The file is not copied to a remote converter as part of opening it.

Typical flow:

1. Open Parquet Fox in a desktop browser.
2. Drop the .parquet file (or pick it from disk).
3. Inspect columns and a few rows, then run SQL if you need a filter.

You get a peek or a filtered export without conda, pip, or a notebook. Limits are device RAM and the WASM memory cap—not a warehouse.

When this is the wrong choice: You need scheduled pipelines, joins across many remote tables, or long-running batch jobs. For those, use Python, DuckDB scripts, or a warehouse—not a tab.

Do not treat an upload-to-a-server “Parquet to CSV” site as the no-Python path when the data is confidential. That is a different trust model: the file leaves your machine.

Other no-Python locals (short)

These still skip Python. They do require installing something.

DuckDB CLI. Install the CLI binary, then query the file in place:

duckdb -c "SELECT * FROM 'data.parquet' LIMIT 20"

No Python package. No notebook. When this is the wrong choice: you cannot install a binary, or you need a clickable grid for someone who will not type SQL.

Desktop Parquet viewer. Install an app, use File → Open, browse rows. When this is the wrong choice: you cannot install desktop software, you need repeatable scripts, or you have not checked whether the app sends samples off-device.

Neither of these is a substitute for a warehouse. Both are local methods. Details and tradeoffs live on How to Read a Parquet File.

When Python still wins

Use Python when you already have it, when the next step is a DataFrame in a pipeline, or when you need scripts in CI. pandas read_parquet is the default there. This page does not repeat that tutorial.

If that is your situation, follow the pandas section on How to Read a Parquet File instead of forcing a browser tab.

What to do next

1. If you cannot install anything, open a file in the browser and confirm you can see rows.
2. If you need to choose among Python, DuckDB, desktop, and browser, use the decision table on How to Read a Parquet File.