Parquet Data Viewer

• Native Mac app for Apache Parquet files
• Columnar format from pandas, Spark, DuckDB, and every data pipeline.
• Adds real data-grid thumbnails and Quick Look, plus a table browser with sort, filter, and CSV export.
• Reads Snappy, Zstandard, gzip, and uncompressed files with flat schemas

Parquet is where data pipelines keep their data. pandas writes it, Spark writes it, DuckDB reads it, and every lakehouse in the world is built on it. On a Mac, a .parquet file is a blank icon, and the usual way to see what is inside is to open a Python session and hope you remember the incantation.

Parquet Viewer opens them like documents. Thumbnails show real rows and columns in your file browser, Quick Look shows the data from the Space bar, and the app gives you a fast grid with sorting, per-column typed filters and CSV export. The columns view shows what is genuinely inside the file: which program wrote it, how many row groups, and for every column its type, compression codec, encodings, null count and both compressed and uncompressed size. Files written by pandas carry their own metadata block, and that is shown too.

Reading Parquet meant implementing the Thrift compact protocol to parse the footer metadata, then a Snappy decompressor by hand, gzip through the system compression framework, and Zstandard through the official BSD-licensed reference decoder because Polars and Iceberg now default to it. On top of that sit the plain and dictionary encodings, the run-length and bit-packed hybrid, definition levels for nulls, and logical types for dates and timestamps.

Scope, stated plainly: flat schemas and data page v1, with Snappy, Zstandard, gzip or no compression.

Requirements

Universal binary, so it runs on both Intel and Apple Silicon Macs.

• macOS 14 (Sonoma) or later
• Intel Mac or Apple Silicon (M1 or newer)

Installation

Download the file, unzip it, and copy it into your Applications folder. Open it once and it registers the Quick Look handler itself. Instructions are included with the download.

Apps downloaded here are signed by Apple, so you shouldn't see any Gatekeeper warnings.

Licence

What you may do: install and use the software, at no cost, on any Mac you own or control, for personal or business purposes.

What you may not do: sell, rent, lease, sublicense or otherwise charge anyone for the software; redistribute, repackage, bundle or host it for download (please point people at the official download page instead); or modify, adapt, translate, reverse-engineer, decompile or disassemble it, except to the limited extent applicable law expressly permits.

The software is licensed, not sold. All rights not expressly granted above are reserved by the copyright holder, who may offer future or paid versions under different terms.

Privacy

100% on-device. No accounts, no tracking, never connects to the internet, collects no data whatsoever.

Feedback

Please contact me with any feedback, issues, suggestions or requests!