Dbt: Flag personally identifiable information in dbt docs

[jgillies]

Feature

Feature description

Add ability to flag PII-containing fields and/or models in dbt documentation

Who will this benefit?

Anyone needing to deal with PII handling responsibilities

[drewbanin]

Thanks for the feature request @jgillies! We currently surface schema tests in the report, though it would be really great to also surface _attributes_ of columns, including (but not limited to) their status as PII.

That interface currently looks like this:

Are you imagining that you'd tag these columns as PII in schema.yml? It might be compelling to add another field to schema.yml for properties that looks something like this:

version: 2

models:
  - name: some_model

    columns:
        - name: customer_name
          properties:
              - pii
          tests:
              - not_null

A property like that could flow through to the documentation website, for instance

[jgillies]

Yep, that's exactly what I had in mind. And expanding the properties beyond just PII is a great idea, too.

edit: another thought - it would be cool to be able to filter on the properties as well, so you could see all columns with a particular property set.

[MartinGuindon]

+1 on this. This would be extremely useful.