{
  "id": 19991835,
  "title": "Fabric PII: What Would You Do When Ingesting Sensitive Data?",
  "url": "https://urgent.news/2026/11/14/fabric-pii-what-would-you-do-when-ingesting-sensitive-data",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-11-14T14:36:03.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/geekindata/fabric-pii-what-would-you-do-when-ingesting-sensitive-data-2b64"
  },
  "original_language": "en",
  "account": "When dealing with sensitive data, it is crucial to have a clear and consistent redaction process in place. This ensures that all personal information, such as names, phone numbers, and email addresses, are appropriately removed or replaced before the data is shared or published. By implementing a config-driven approach, the redaction rules can be easily managed and reviewed by relevant stakeholders. This includes specifying which columns contain sensitive data, the patterns to match for redaction, and the replacement tokens to use. By centralizing the redaction configuration in a single cell or module, it becomes easier to maintain and ensure that the same rules are applied consistently across different notebooks or exports. Additionally, the redaction process should be integrated with the publish gate, meaning that the data cannot be published unless the redaction step is successful. This helps prevent accidental leakage of sensitive information and provides an additional layer of security. It is also important to consider the different classifications of data, such as personal versus operational, and handle them appropriately based on the organization's policies. By separating the allowed versus forbidden categories and providing examples, the redaction rules become more transparent and easier to understand. Furthermore, the use of stable tokens, such as {{EMAIL}} or {{PHONE}}, allows for easier auditing and counting of redacted data. This is particularly useful when reporting on the number of sensitive fields that have been removed. Known-name lists or UDFs (user-defined functions) can also be incorporated as an additional layer of redaction, helping to identify and handle informal names or titles that may not be caught by regular expressions. Overall, a well-defined and documented redaction contract, embedded within the notebook or close to the notebook that owns the export path, ensures that the privacy controls are visible and auditable. By treating redaction as a gate rather than an afterthought, the integrity of the data is maintained, and the likelihood of privacy breaches is minimized.",
  "summary": "I used AI tools to help draft and edit this post. I checked the facts and the wording, and the original lives at geekindata.com . The diagrams are on that page. Redaction as a publish gate Key takeaways All lakehouse-leaving extracts use a shared redaction config module or notebook cell contract. Allowed versus forbidden free-text categories are written with examples. Replacement tokens are…",
  "key_points": [
    "Implement a config-driven redaction approach with clear rules for sensitive data.",
    "Integrate redaction with publish gate to prevent accidental data leakage.",
    "Use stable tokens and known-name lists for easier auditing and counting of redacted data."
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}