Parquet File Viewer
Open a .parquet file and inspect its schema, metadata, and rows —
no Spark, no Python, no upload. Great for a quick sanity check on a file
before you load it into a lakehouse.
Drag & drop a .parquet file here, or
File metadata
Schema
About this Parquet viewer
Apache Parquet is a columnar file format used across the modern data stack —
Spark, Microsoft Fabric, Databricks, DuckDB, pandas, and every major lakehouse.
The catch is that Parquet is binary, so you can't just open a .parquet
file in a text editor to see what's inside. This tool lets you
view a Parquet file online without uploading it: pick a file and it's
parsed entirely in your browser using the open-source
hyparquet
library. You get the row count, row groups, the tool that wrote the file, the
compression codec, a full schema tree with physical and logical types, per-column
null counts, and a paginated preview of the actual rows.
It's built for data engineers, analytics engineers, and QA engineers who need a fast answer to "what's actually in this file?" — checking column names and types before building a pipeline, confirming a Bronze-layer export, or debugging a schema-drift bug — without spinning up a Spark session or writing a throwaway pandas script. Snappy- and ZSTD-compressed files are supported.
FAQ
Is my Parquet file uploaded to a server?
No. The file is read and decoded locally by JavaScript in your own browser tab. Nothing is sent anywhere — there's no server that receives your data, and the tool works offline once the page has loaded.
How do I view a Parquet file online without Python or Spark?
Just drop the .parquet file onto this page (or click "Choose file"). It reads the footer metadata and rows directly in the browser, so you don't need pandas, PyArrow, DuckDB, or a Spark cluster to peek inside a file.
Does it support Snappy and ZSTD compression?
Yes. Snappy is the most common Parquet codec and is fully supported, along with ZSTD, GZIP, and uncompressed files, via the hyparquet-compressors package that also runs client-side.
Can it handle large Parquet files?
You can open large files, but to keep your browser responsive the preview loads 100 rows at a time and caps the on-screen table at about 1,000 rows. Metadata and schema are always read in full. For multi-gigabyte files, a desktop tool will be more comfortable.
Independent community tool — not affiliated with the Apache Software Foundation or the Parquet project. Powered by the open-source hyparquet library.