Preserve raw JSON states when building a spreadsheet view

A blank spreadsheet cell can hide several states

Preserve missing, unsupported and unresolved states so a friendlier format does not make uncertain data look finished.

Before paying for technical-document extraction, inspect how the source meaning survives in the output. ParseMill’s public material-register example gives you a synthetic PDF and downloadable evidence to compare. Bring your own document type and acceptance checks to a scope discussion; the self-service invoice plan is a separate offer.

Do not infer business meaning from JSON type alone. Null may represent unknown or unavailable information under a particular schema, but the schema or source must establish that meaning. An absent property may indicate a different processing path rather than a reported blank.

Use a JSON-to-view mapping table

The objects are synthetic examples of one field, not complete extraction records. A real mapping should also retain stable row identity, units and source references. Avoid collapsing all nonnumeric cases into zero merely to satisfy a numeric worksheet column.

RFC 8259: JSON data interchange defines distinct JSON value types. Preserve those distinctions through a declared view mapping rather than relying on spreadsheet defaults.

Build the view deliberately

1. Read the schema and document the meaning of each field and missing state. Identify which properties are required and which may be absent.
2. Create separate raw representation, typed value, state and review-status columns where the distinction matters. Keep a source record key.
3. Flatten arrays and nested objects using an explicit row model. Reconcile generated rows to their parent records so repeated observations do not merge silently.
4. Export and reopen the view, then compare the fixture states with the raw JSON. Confirm that filtering or formulas do not treat every blank-looking cell identically.

JSON Schema: Object validation explains required properties and object validation. A required property’s absence is a different schema question from a present property whose value is null.

Handle convenience transformations visibly

A display label such as “Not available” can help readers, but retain the original state it represents. If several raw states share one label, include a separate state code so they remain distinguishable for review.

Do not use a successful JSON parse as proof of a complete extraction. Syntax validity establishes that the data can be read, while schema validation and source review address structure and meaning separately.

Check the view against the raw record

The fixture passes when all five cases remain distinguishable and the qualified string retains its comparator. A reviewer should be able to reconstruct the raw state from the view and its linked record.

Use this raw-state fixture to describe the JSON integration you need in a scoped workflow discussion. Agree the schema, missing-state mapping and source-review requirements before a wider pilot. The current self-service invoice workspace delivers an Excel workbook; it does not promise this general integration.