ELT-Bench

该数据集名为ELT-Bench，是一个端到端的基准测试，旨在评估人工智能代理构建ELT（提取、转换、加载）管道的能力。该数据集包含100个管道，涵盖了包括835个源表和203个数据模型，这些数据模型覆盖了多个领域。此外，该基准测试还评估人工智能代理在处理复杂的数据工程工作流程方面的能力，这些工作流程涉及与数据库的交互、编写代码和SQL查询，以及协调每个管道阶段的工作。具体规模包括100个管道、835个源表和203个数据模型，任务则是评估人工智能代理在构建ELT管道方面的能力。

This dataset, named ELT-Bench, is an end-to-end benchmark designed to evaluate the capability of AI Agents in constructing ELT (Extract, Transform, Load) pipelines. It comprises 100 pipelines, covering 835 source tables and 203 data models that span multiple domains. Additionally, this benchmark assesses the ability of AI Agents to handle complex data engineering workflows, which involve interactions with databases, writing code and SQL queries, and coordinating work across each pipeline stage. The specific scale of this benchmark includes 100 pipelines, 835 source tables and 203 data models, and its core task is to evaluate the capability of AI Agents in building ELT pipelines.