Terminal
Terminal/Skill
Load files, object storage, dataframes, or external databases into MotherDuck using an appropriate bulk ingestion path.
By MotherDuck
Works with
npx skills add motherduckdb/agent-skills --skill motherduck-load-dataCREATE SECRET and cloud-storage docs for protected-object-store workflows..duckdb, .ddb, or attached DuckDB database.COPY, CTAS, or INSERT ... SELECT over row-by-row insertsCREATE TABLE AS SELECT for first loads and INSERT INTO ... SELECT for appends.CREATE OR REPLACE DATABASE remote_name FROM CURRENT_DATABASE(), an attached local database, a local file path from a native client, or a remote .duckdb file URL such as S3. Remote and local file imports physically copy data into MotherDuck; database/share clone sources are zero-copy.CREATE DATABASE ... FROM '<cloud-url>' with the required cloud secretFor answer, review, or planning requests, recommend the loading path without mutating data. For load or implementation requests, perform the requested in-scope write and validation; ask before destructive replacement or a broader external write.
Read only the reference sections needed for the current task.
references/INGESTION_PATTERNS.md for format-specific options, cloud-storage secrets, Postgres-endpoint loading tradeoffs, Python dataframe paths, and advanced ingestion patterns.Load related skills only for missing capabilities; reuse established context.
motherduck-connect for choosing between the Postgres endpoint and a DuckDB client pathmotherduck-explore for inspecting destination databases and validating landed tablesmotherduck-query for writing CTAS, append, and validation SQLmotherduck-model-data for promoting landed data into staging and analytics tablesmotherduck-ducklake only when object-storage-backed lakehouse storage is an explicit requirementmotherduck-cli when a shell-based load should stream structured output to files