duckdb

name	duckdb
description	Analytical SQL queries on local files via DuckDB. Use for querying CSV, Parquet, JSON, and Excel files with SQL, aggregating data, joining datasets, and exporting results — all without a database server.
metadata	{"version":"1.0.0","displayName":"DuckDB","author":"gremlin","category":"data","icon":"database","tags":["sql","data","analytics","csv","parquet","json"],"install":"which duckdb || pip install duckdb-cli\n","allowedCommands":["duckdb"]}

DuckDB

You have access to duckdb for running SQL queries directly on files.

Usage

# Interactive mode
duckdb

# Query a file directly
duckdb -c "SELECT * FROM 'data.csv' LIMIT 10"
duckdb -c "SELECT * FROM 'data.parquet'"
duckdb -c "SELECT * FROM read_json_auto('data.json')"

# Persist a database
duckdb mydb.duckdb -c "CREATE TABLE t AS SELECT * FROM 'data.csv'"

Tips

• Query files directly without import: SELECT * FROM 'file.csv' — DuckDB auto-detects format.
• Glob patterns: SELECT * FROM 'logs/*.csv' to query multiple files at once.
• Remote files: SELECT * FROM 'https://example.com/data.csv' works out of the box.
• Export results: COPY (SELECT ...) TO 'output.csv' (HEADER, DELIMITER ',') or .parquet.
• Inspect schema: DESCRIBE SELECT * FROM 'file.csv'
• Summary stats: SUMMARIZE SELECT * FROM 'file.csv'
• Use -json flag for JSON output: duckdb -json -c "SELECT ..."
• Use -markdown flag for markdown table output.
• Use -csv flag for CSV output.
• DuckDB supports window functions, CTEs, PIVOT/UNPIVOT, QUALIFY, list/struct types, and regex.
• For large files, DuckDB streams and doesn't load everything into memory.
• Read Excel: SELECT * FROM st_read('file.xlsx') (requires spatial extension: INSTALL spatial; LOAD spatial;).
• Multiple statements: pipe a .sql file with duckdb < queries.sql.