What is a Parquet file?

Parquet is a compact file format used for large data exports. It stores column names and types with the data. Because it is binary, you cannot read it in a text editor like a CSV.

Do I need Python or DuckDB to open a Parquet file?

No. Drop a .parquet file up to 50 MB into the viewer above to open it in your browser. Python and DuckDB are also useful when you want to work in code.

How big a Parquet file can I open?

The browser viewer accepts .parquet files up to 50 MB. It uses virtual scrolling to display every parsed row. Performance depends on the file, browser, and device.

Is my Parquet file uploaded anywhere?

Your file contents are parsed in your browser and are not uploaded to Loam. File type, size, or error details may still be recorded for analytics and error reporting.

Can I ask questions about the data in a Parquet file?

The free chat tool accepts CSV and TSV, not Parquet. To keep Parquet data connected to your AI, start with Loam and upload the file separately.

How to open a Parquet file and inspect the real rows

Parquet is a binary file format, so a text editor cannot display it as a table. Drop a .parquet file up to 50 MB here to inspect rows and column details, sort, filter, and download the active view as CSV. The file contents are parsed in your browser and are not uploaded to Loam.

Drop a .parquet file here or click to choose a file

Your file contents are parsed in your browser and are not uploaded to Loam.

See every rowCheck column detailsSort and filter rowsDownload your view

Open it in your browser

Drop a .parquet file up to 50 MB into the viewer above. It displays every parsed row and useful column details, including types and compression. You can sort, filter, and download the active view as CSV. The file contents are not uploaded to Loam.

If you'd rather open it in code

If you already have Python or DuckDB set up, two lines will read a Parquet file. Reach for these when you're scripting; reach for the viewer above when you just want to look.

Python (pandas)

import pandas as pd
df = pd.read_parquet("data.parquet")
print(df.head(20))

DuckDB

SELECT * FROM 'data.parquet' LIMIT 20;

Both options require an installed environment. Use the Parquet viewer when you only need to inspect the file.

After you inspect the file

Sort and filter the rows, then download the active view as CSV. To keep Parquet data available to your AI, start with Loam and upload the file separately. The browser viewer does not transfer the selected file. If you need a plain text copy for Excel or Google Sheets, use the Parquet to CSV converter. That converter uploads the file separately to Loam temporary storage. Working with another format? Open a CSV file the same way.

Frequently asked questions

Ready to ground your AI in real data?

Start with Loam to upload Parquet data and connect your AI. This browser viewer does not transfer the selected file automatically.

Start with Loam Open the Parquet viewer