Enterprise LayerEnterprise Layer
FDE 进入企业现场，经常先面对一堆 Excel、CSV、ERP 导出和口径不一致的指标。本章不是培养数据工程师，而是训练学员会用 AI 工具快速整理、解释和验证数据。In the field, an FDE often starts with Excel files, CSV exports, ERP dumps, and conflicting metric definitions. This chapter does not train data engineers; it trains learners to use AI tools to organize, explain, and verify data.
学习产出Learning Outcomes
01
用 AI 工具生成字段字典和数据盘点表Use AI tools to generate field dictionaries and data inventories
02
让 AI 辅助发现缺失、重复、异常和口径冲突Ask AI to find missing values, duplicates, anomalies, and metric conflicts
03
理解 SQL/脚本的基本用途但不做语法专精Understand the purpose of SQL/scripts without deep syntax study
04
输出业务可读的数据说明和风险提示Ship business-readable data notes and risk warnings
扩展学习资源Resources
Cursor Docs 用 IDE Agent 生成数据清洗脚本并解释结果。 Use an IDE agent to generate data-cleaning scripts and explain results. DuckDB Docs 只作为本地 CSV/Excel 分析工具概念参考，不要求掌握数据库工程。 Use as a local CSV/Excel analysis reference, not as database engineering training. Select Star SQL 需要看懂 AI 生成 SQL 时再补基础。 Use it when you need to understand AI-generated SQL basics.
检查清单Checklist
数据与安全Data and Security
- 数据分级是什么？是否包含个人信息、客户资料、合同、财务或敏感业务数据？How is the data classified? Does it include personal, customer, contract, finance, or sensitive business data?
- 数据如何进入系统：本地 Excel/CSV、ERP 导出、数据库只读账号、飞书表格还是 API？How does data enter the system: local Excel/CSV, ERP export, read-only database account, Feishu table, or API?
- 是否需要脱敏、最小权限、访问日志和人工复核？Do we need masking, least privilege, access logs, and human review?