Scholarlink

The Business Intelligence process involves collecting, preparing, and transforming data from various sources, which is then stored in a Data Warehouse (DW) for analysis. Given that the Extract-Transform-Load (ETL) phase is crucial for data quality, we argue that provenance collection can enhance ETL processes by tracking transformations, allowing for subsequent inspection and auditing. Accordingly, we propose ProvETL, a solution that integrates provenance management into ETL routines, providing data-driven debugging and accountability for data transformations. We conducted several evaluations in a real-world DW of a Brazilian University that indicated ProvETL can effectively spot data inconsistencies and audit their related transformations.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enriching ETL with Provenance Data

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Matheus Vieira,
  • Thiago de Oliveira,
  • Leandro Cicco,
  • Daniel de Oliveira,
  • Marcos Bedo

摘要

The Business Intelligence process involves collecting, preparing, and transforming data from various sources, which is then stored in a Data Warehouse (DW) for analysis. Given that the Extract-Transform-Load (ETL) phase is crucial for data quality, we argue that provenance collection can enhance ETL processes by tracking transformations, allowing for subsequent inspection and auditing. Accordingly, we propose ProvETL, a solution that integrates provenance management into ETL routines, providing data-driven debugging and accountability for data transformations. We conducted several evaluations in a real-world DW of a Brazilian University that indicated ProvETL can effectively spot data inconsistencies and audit their related transformations.