DuckDb

The DuckDbTools toolkit enables an Agent to run SQL and analyze data using DuckDb.

DuckDbTools enable an Agent to run SQL and analyze data using DuckDb.

Prerequisites

The following example requires DuckDB library. To install DuckDB, run the following command:

1uv pip install duckdb

For more installation options, please refer to DuckDB documentation.

Example

The following agent will analyze the movies file using SQL and return the result.

1from kern.agent import Agent
2from kern.tools.duckdb import DuckDbTools
3
4agent = Agent(
5    tools=[DuckDbTools()],
6    system_message="Use this file for Movies data: https://kern-public.s3.amazonaws.com/demo_data/IMDB-Movie-Data.csv",
7)
8
9agent.print_response("What is the average rating of movies?", markdown=True, stream=False)

Toolkit Params

Parameter	Type	Default	Description
`db_path`	`str`	`None`	Specifies the path to the database file.
`connection`	`DuckDBPyConnection`	`None`	Provides an existing DuckDB connection object.
`init_commands`	`List`	`None`	A list of initial SQL commands to run on database connection.
`read_only`	`bool`	`False`	Configures the database connection to be read-only.
`config`	`dict`	`None`	Configuration options for the database connection.

Toolkit Functions

Function	Description
`show_tables`	Function to show tables in the database
`describe_table`	Function to describe a table
`inspect_query`	Function to inspect a query and return the query plan. Always inspect your query before running them.
`run_query`	Function that runs a query and returns the result.
`summarize_table`	Function to compute a number of aggregates over a table. The function launches a query that computes a number of aggregates over all columns, including min, max, avg, std and approx_unique.
`get_table_name_from_path`	Get the table name from a path
`create_table_from_path`	Creates a table from a path
`export_table_to_path`	Save a table in a desired format (default: parquet). If the path is provided, the table will be saved under that path. Eg: If path is /tmp, the table will be saved as /tmp/table.parquet. Otherwise it will be saved in the current directory
`load_local_path_to_table`	Load a local file into duckdb
`load_local_csv_to_table`	Load a local CSV file into duckdb
`load_s3_path_to_table`	Load a file from S3 into duckdb
`load_s3_csv_to_table`	Load a CSV file from S3 into duckdb
`create_fts_index`	Create a full text search index on a table
`full_text_search`	Full text Search in a table column for a specific text/keyword

You can use include_tools or exclude_tools to modify the list of tools the agent has access to. Learn more about selecting tools.

Developer Resources

View Tools