datadata-dql

datadata-dql is a skill for AI agents to write DQL (Data Query Language) scripts. DQL is a Starlark-based extended scripting language supporting data transformation, HTTP requests, 2D drawing, and more.

Features

• DataFrame / Series Operations — Data transformation, cleaning, aggregation, merging, anomaly detection
• SQL Queries — Execute SQL via query() and return a DataFrame
• HTTP Requests — Call external APIs via fetch()
• 2D Drawing — Canvas basic shape drawing
• Built-in Modules — JSON, time, math, etc.

DQL Script Basics

A DQL script is essentially a function body and must have a return statement to return data:

df = query("SELECT * FROM users LIMIT 100")
return df

All DQL extension functions are built-in global names — no import needed:

• DataFrame, Series — Data structures
• query() — SQL query
• fetch() — HTTP request
• concat() — Data concatenation
• json, math, time — Built-in modules

Full Language Reference

For the complete DQL syntax, API signatures, and examples, see the DQL Query Scripts chapter.

Use Cases

Scenario	Example
Data Cleaning	Filter, type-cast, and handle missing values in query results
Data Transformation	Column calculations, group aggregation, multi-table merging
Data Generation	Fetch external data via HTTP API and convert to DataFrame
Custom Analysis	Write complex data processing pipelines
Visualization	Draw charts using Canvas