How should I handle information that is “not applicable” rather than missing?

Empty is ambiguous: it might mean “not collected yet” or “does not apply.” Data-quality checks should only flag genuine omissions. In Busabase, separate applicability from the value.

Pattern

Examples:

• Physical delivery → applicable = Applicable, shipping address required.
• Online service → applicable = Not applicable, leave the address empty on purpose.

*Diagram · applicability first; DQ only when required and empty*

*Diagram · Select options: Applicable / Not applicable*

Data-quality rule

Flag a problem only when:

applicable = Applicable (or true) AND value is empty.

Do not treat every blank as an error. Do not use blank alone to mean N/A—always set applicability explicitly.

How to set up views / checks

1. Needs attention: applicable = Yes AND value empty.
2. N/A set: applicable = No (optional audit view).
3. Forms / intake: ask applicability first; hide or skip the value when N/A if your UI allows, but still store the applicability choice.

How to verify

• Online-service row with Not applicable and empty address → not in the DQ failure view.
• Physical row with Applicable and empty address → is flagged.
• A row that never set applicability → treat as incomplete process (prompt to choose), not as silent N/A.

Boundaries

• Don’t invent placeholder values (“N/A” string inside the address field) if you already have an applicability field—that confuses exports.
• Multi-field groups (billing vs shipping): each group can have its own applicability.
• Historical imports: backfill applicability where you can; until then, keep DQ scoped to records that have the gate filled.

If you tell us which fields are conditional (shipping, tax ID, device serial, …), we can sketch a minimal field pair list for your Base.