Sample Parquet files

Apache Parquet is a columnar data storage format, a standard in data engineering: Spark, Pandas, DuckDB, BigQuery. Below are sample .parquet files with various schemas, types and compression options to download

Sample file: Simple Parquet: 5 rows, basic types
Simple Parquet: 5 rows, basic types
1.2 Kb
↓
Sample file: Users: 1000 rows, realistic dataset
Users: 1000 rows, realistic dataset
35 Kb
↓
Sample file: All data types: int, float, bool, string, binary, date, timestamp, decimal
All data types: int, float, bool, string, binary, date, timestamp, decimal
4.2 Kb
↓
Sample file: Nested data: struct columns and lists (list)
Nested data: struct columns and lists (list)
2.5 Kb
↓
Sample file: Large Parquet: 500,000 rows (~5 MB)
Large Parquet: 500,000 rows (~5 MB)
4.9 Mb
↓
Sample file: Empty Parquet: schema without rows
Empty Parquet: schema without rows
621 bytes
↓
Sample file: gzip compression: 20,000 rows
gzip compression: 20,000 rows
308 Kb
↓
Sample file: Uncompressed: the same 20,000 rows
Uncompressed: the same 20,000 rows
1.4 Mb
↓

🧠 How Parquet works

Parquet stores data by column, not by row. Values of one column sit next to each other, so they compress better and can be read selectively — you can pull only the columns you need. The file is split into row groups, and the footer at the end holds the schema and statistics (min/max) for quickly skipping blocks.

📊 Table rows × columns 📑 Columnar data by column 🗜️ Encoding + compression 📦 row groups + footer 💾 .parquet ready file