Parquet Import

Read Parquet files directly in SQL from HTTP URLs or the embedded filesystem.

Installation

SQL

CREATE EXTENSION parquet;

Read from HTTP URL

SQL

-- Read Parquet file from URL
SELECT * FROM read_parquet('https://example.com/data.parquet');

-- With column selection
SELECT col1, col2 FROM read_parquet('https://example.com/data.parquet');

Read from fs9 Filesystem

SQL

-- Read Parquet file from embedded filesystem
SELECT * FROM read_parquet('/data/analytics.parquet');

COPY FROM with Parquet

SQL

-- Import Parquet data into existing table
COPY my_table FROM '/data/import.parquet' FORMAT parquet;

-- Import from URL
COPY my_table FROM 'https://example.com/data.parquet' FORMAT parquet;

Common Patterns

SQL

-- Combine with aggregations
SELECT
  category,
  COUNT(*) as count,
  AVG(value) as avg_value
FROM read_parquet('/data/sales.parquet')
GROUP BY category;

-- Filter while reading
SELECT * FROM read_parquet('/data/logs.parquet')
WHERE timestamp > '2026-01-01';