How to Open Parquet Files Without Spark or Python
Apache Parquet is the de-facto storage format for big data pipelines. But if someone sends you a .parquet file and you need to quickly check its contents or schema, firing up a Spark cluster or setting up a Python environment feels like overkill.
Here are three practical ways to inspect Parquet files with zero infrastructure.
Option 1: Use a browser-based Parquet viewer (fastest)
The fastest option is a tool that reads the file entirely in your browser using WebAssembly — no install, no upload, no waiting.
Hexabench Parquet Viewer lets you drag and drop a .parquet file and immediately see the rows, column types, null counts, and sample values. Your file never leaves your machine.
Use this when you need a quick sanity check on a file from a colleague or an ETL output.
Option 2: DuckDB (best for SQL queries)
If you have DuckDB installed locally (brew install duckdb on macOS), you can query Parquet files directly with SQL:
-- Show first 10 rows
SELECT * FROM read_parquet('your_file.parquet') LIMIT 10;
-- Inspect schema
DESCRIBE SELECT * FROM read_parquet('your_file.parquet');
-- Aggregate
SELECT status, COUNT(*) FROM read_parquet('your_file.parquet') GROUP BY status;DuckDB is fast, handles large files well, and supports the full Parquet spec including nested types and dictionary encoding.
Option 3: pandas (if Python is already available)
If you have Python with pandas and pyarrow installed:
import pandas as pd
df = pd.read_parquet('your_file.parquet')
print(df.head())
print(df.dtypes)
print(df.isnull().sum())This works well for exploration but requires pip install pandas pyarrow first. If the file is very large, DuckDB or the browser viewer will be faster for initial inspection.
Which option to use?
- Quick inspection — browser-based viewer, zero setup
- SQL queries and filtering — DuckDB CLI or browser SQL query tool
- Python data pipeline — pandas + pyarrow