DuckDB 2 cheatsheet

DuckDB 2: Files and command line

6 common DuckDB 2 tasks.

Write query results to Parquet

Export a result as a compressed Parquet file.

COPY (SELECT * FROM sales) TO 'sales.parquet' (FORMAT parquet);

Export a result as CSV with a header.

COPY (SELECT * FROM sales) TO 'sales.csv' (HEADER, DELIMITER ',');

Start the interactive shell on a database file. Omit the file for an in-memory database.

duckdb analytics.duckdb

Run one statement and exit, which suits scripts.

duckdb -c "SELECT count(*) FROM read_parquet('data/*.parquet')"

Print results as markdown, CSV or JSON from the CLI.

.mode markdown

Load the httpfs extension, then read remote files like local ones.

INSTALL httpfs;
LOAD httpfs;
SELECT count(*) FROM read_parquet('https://example.com/data.parquet');