Query Parquet files with SQL

ParquetView’s SQL tab is DuckDB running as WebAssembly. The dialect is DuckDB’s, not a toy subset: joins, CTEs, window functions, and GROUP BY work. Queries run on your machine against the table created from the file you opened.

Get started

1. Open a Parquet file.
2. Switch to the SQL tab. Table-name chips above the editor copy the identifier you should use in FROM.
3. Press Ctrl+Enter or Cmd+Enter.

Examples

-- first look
SELECT * FROM my_file LIMIT 100;

-- distribution
SELECT category, COUNT(*) AS n
FROM my_file
GROUP BY category
ORDER BY n DESC;

-- filter then aggregate
SELECT region, AVG(sales) AS avg_sales
FROM my_file
WHERE year >= 2023
GROUP BY region;

After a query, the grid shows the result. Export that result to CSV or JSON. Query history stays in this browser; it is not synced to an account.

Joining two files

Load both files, then join in SQL, or use the Combine dialog (inner / left / right / full). Give the result a name and it appears in the library like any other table.

Natural language (optional)

“Ask your data” can draft SQL with Gemini if you provide an API key. That request goes to Google’s API from your browser, not through a ParquetView backend. Leave it off if the file is sensitive. Details are on the privacy page.

Related

• Open a Parquet file
• Export query results to CSV