Parquet X-ray

Parquet X-ray

Lets you inspect how a Parquet file is built inside, from row groups and column chunks to pages, indexes, bloom filters and the footer. Runs in the browser, for data and ML engineers.

Made by cfahlgren1 on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: Parquet X-ray is a browser tool for looking inside a Parquet file. It draws how the file is stored: the row groups, their column chunks and pages, plus the indexes, the bloom filters and the footer. You can open a page's details in a popover to see how the bytes are organized.

Who it suits: It suits people who work with Parquet files and want to see why a file is shaped the way it is, or learn the format.

Industry | Who uses it | How they use it | Example
Data engineering | Data engineers | Check how row groups and pages are laid out in a file | An engineer opens a file to see how its columns are chunked
Machine learning | ML engineers | Inspect a dataset file from the Hub before loading it | An engineer looks at the footer of a Hub dataset file
Analytics | Analysts | Find out what indexes and bloom filters a file carries | An analyst checks whether a file has page indexes
Education | Students | Learn the Parquet format by looking at a real file | A student opens a file and follows it from footer to page

You can skip it if you only read Parquet data through a library and never need the file's layout.

Where to use it: You use it in these places.

• A web browser, on the Space's page on Hugging Face
• A local copy of the built page, opened in a browser on Windows, macOS or Linux
• A development server, started with npm, for people who change the code

Use case | Who | What they get back
Inspect a remote file | Data and ML engineers | The file's layout, read from a URL without downloading all of it
Inspect a local file | Anyone | The same layout view for a file you open from disk
Study the format | Students | A visual tour of row groups, column chunks and pages

How to set it up: Online use needs only a browser. A local copy needs Python for the download step. Changing the code needs npm.

• A web browser
• Python with pip, only to fetch a local copy
• Node and npm, only to develop the project

1. Open the Space's page on Hugging Face to use it online, with nothing to install.
2. For a local copy, install the Hugging Face Hub package.
3. Download the Space's files into a folder and move into it.
4. Load dist/index.html in your browser.

pip install -U huggingface_hub
hf download cfahlgren1/parquet-xray --repo-type space --local-dir parquet-xray
cd parquet-xray

To work on the code, the README gives these development commands.

npm install
npm run dev
npm run verify

How to use it: Give the page the location of a Parquet file. That can be an address on the Hub, an hf:// path, a web address that permits CORS range requests, or a file you pick from your disk. The page reads only the footer and the page indexes, so a 2 GB file needs about 3 MB downloaded before it opens.

Tool | What it does
npm install | Installs the project's dependencies
npm run dev | Starts the development server
npm run verify | Runs the lint, the type check, the tests and the build

Example: You paste the address of a Parquet file on the Hub. The page reads its footer and indexes, then shows the row groups. You open one to see its column chunks and pages, and click a page to see its details in a popover.

For newcomers and experts: A Parquet file stores a table in blocks called row groups, and each column inside a block is a chunk made of pages. This page draws that structure. Try it by pasting the address of a file you know and opening the first row group.

Under the hood, the page uses range requests to read just the footer and page indexes, which is why a large file opens fast. Parsing uses the hyparquet library. The repository's config files point to Svelte, Vite and Playwright, and CONTRIBUTING.md explains how the code is organized.

Paste a URL or open a local file
Only the footer and page indexes are read
hyparquet parses the file's layout
Row groups, column chunks and pages are shown

Good to know: Remote files must allow the kind of requests the page makes.

• License: Apache 2.0, as the README states.
• Cost to run: none named.
• Data it reaches: the Parquet files you point it at, by URL or from your disk. It reads only the footer and page indexes of a remote file.
• Not for: URLs that do not allow CORS range requests.

On Hugging Face, 13 November 2026

• Likes: 7
• Built with: static
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/cfahlgren1/parquet-xray
• Files: https://huggingface.co/spaces/cfahlgren1/parquet-xray/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.