Columnar / File formats / Parquet

One file. Many layers.

Parquet organizes data into columns so readers can select the pieces they need. Explore how its schema, row groups, pages and encodings work together.

Start the guide →Explore a file

01 / THE EXAMPLE

Parquet, Explained

A guided tour of a 3,000-byte file containing 344 penguins. Every displayed measurement comes from the example.

Follow the bytes →

02 / YOUR FILE

Read its anatomy

Replay the guide's concepts on a local file, reading the footer first and fetching detail on demand.

Explain your file →

03 / THE READER

TypeScript API

Metadata, column reads, nested records and statistics, with explicit value types and supported features.

Inspect the API →

Local inspection

The explorer does not upload your files. It decodes selected column chunks in a browser worker and shows a page of values at a time. Large chunks and exports can still use substantial memory.

Open the explorer →

What you can explore

Real bytes, encodings, nested values, statistics and local file inspection.

Compare writer layoutsLive browser execution
From pages to valuesOriginal bytes
Inspect your own fileLive browser execution

The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.