Confirm the schema and row count
Use the metadata panel for the loaded schema, then run a count. A renamed column, unexpected type, or surprising row total is often the fastest signal that you received the wrong file or pipeline version.
Open Parquet
Open a Parquet file online
Apache Parquet is a compressed, columnar file format — which is exactly why double-clicking a .parquet file gets you nothing. It is binary, so Notepad, Excel, and text editors show gibberish, and most advice online starts with "install Python and pandas".
You do not need any of that for a quick look. viewparquet opens local Parquet files directly in your browser: drag the file in and DuckDB-WASM reads it on your device. The file is not uploaded to viewparquet, there is no account, and no desktop software is required.
Updated August 21, 2026 · Reproducible checks, limitations, and primary references are included below.
Open your Parquet fileOpen a Parquet file in three steps
Go to viewparquet.com
No sign-up or install. Core file viewing and SQL processing run in your browser tab.
Drag in your .parquet file
Or click "Open a file" and pick it. DuckDB-WASM parses the file on-device for viewing and SQL; its contents are not uploaded to viewparquet. You can also open a file straight from S3 or a URL.
Browse rows and inspect columns
Browse the virtualized grid, check column names and DuckDB types, and use the table metadata panel for a schema snapshot. Run SQL or export to Parquet, CSV, or JSON Lines when you need more.
A five-minute Parquet file triage
Opening a file proves that it is readable; it does not prove that the rows are complete, unique, or safe to use. Check the structure and a few invariants before exporting or sharing a result.
Test the key fields you actually rely on
Check nulls, duplicate identifiers, timestamp ranges, and category values for the columns that drive the next decision. A grid preview can look plausible while a key field is incomplete farther down the file.
Query narrowly before a large export
Select only the columns and rows needed for the task. This makes mistakes easier to spot and reduces browser work, but it is not a guarantee about memory use or remote bytes transferred.
Reproduce a known result with the bundled sample
Open the flights sample from the homepage and run this query. The checked-in file has 231,083 rows and three columns: delay, distance, and time.
SELECT
count(*) AS rows,
min(delay) AS min_delay,
max(delay) AS max_delay
FROM data;What to expect: Expected for the repository sample verified on August 19, 2026: 231,083 rows, a minimum delay of -58, and a maximum delay of 180. A different result means the input or sample version changed; it is not a universal Parquet expectation.
Primary references
These sources support the format and tool behavior described on this page. Product-specific boundaries are checked against the shipped viewer using the claim-verification method.
Common questions
What app opens a .parquet file?
No default desktop app opens .parquet files — the format is a binary, columnar data file, not a document. Use a Parquet viewer: viewparquet opens the file in your browser with no install, and command-line options include the DuckDB CLI and parquet-tools.
How do I open a Parquet file without Python or Spark?
Open it in a browser-based viewer. Drag the .parquet file into viewparquet and DuckDB-WASM reads it locally — no Python, pandas, Spark, or install required. You can browse rows, inspect column types, and run SQL.
How do I open a Parquet file on Windows?
Open your browser (Edge or Chrome both work), go to viewparquet.com, and drag the .parquet file into the page. Nothing is installed and the file is processed locally. On a managed corporate machine, the site and WebAssembly runtime must still be permitted by your organization policy.
Can Excel open a Parquet file?
No. Microsoft's current Power Query connector matrix marks Parquet as unsupported in Excel. Open the file in viewparquet, export the rows you need to CSV, and import that CSV with Excel's text/CSV workflow instead.
Is it safe to open a confidential Parquet file this way?
For viewing, SQL, and export, DuckDB-WASM processes the local file in your browser and does not upload the dataset to viewparquet servers. Optional AI is a separate path: when you invoke it, your browser sends your message and configured context directly to the chosen provider, and settings can allow samples or bounded query-result values. Review Settings → AI and the provider policy before using AI with confidential data.
What about very large Parquet files?
There is no universal file-size guarantee. Browser memory, Parquet layout, column types, and the query or export all affect the practical limit; the grid reads pages of results, so selecting only the columns and rows you need is the safest approach for a large file.
Continue with a related task
Try it on your own file
Drop a Parquet, GeoParquet, CSV, or JSON file into viewparquet: browse rows, run DuckDB SQL, and export results in-browser without uploading the dataset to viewparquet. Optional AI sends messages and applicable context directly to the provider you choose; review Settings → AI before use.