Hello,
I am writing on behalf of Unito. We are a product that provides deep, two-way sync across work tools. An omission in the API table model prevents us from properly handling dependencies.
Summary: `dateDependencySettings` is documented as part of the table model response. It should carry everything needed to handle date dependencies. Which fields are the start date, end date and duration, which field holds the dependency relation, and the rescheduling settings. In practice the property is never returned in the table model response, so there is no way to read a table's date-dependency configuration through the API.
Setup: a base on a plan with date dependencies, one table with Start date, End date, a duration field, and a self-linking predecessor field.
What I did:
1. PATCH /v0/meta/bases/{baseId}/tables/{tableId} with:
{
"dateDependencySettings": {
"isEnabled": true,
"startDateFieldId": "fldSTART",
"endDateFieldId": "fldEND",
"durationFieldId": "fldDURATION",
"predecessorFieldId": "fldPRED",
"reschedulingMode": "flexible",
"shouldSkipWeekendsAndHolidays": false,
"holidays": []
}
}→ **200**, and the response body echoes the full `dateDependencySettings` object.
2. The Airtable UI's Date dependencies panel for that table now shows all four fields populated exactly as sent — so the write reached the real store, it isn't a no-op.
3. GET /v0/meta/bases/{baseId}/tables for the same table → the table object contains only `id`, `name`, `primaryFieldId`, `fields`, `views`. No `dateDependencySettings`, even though it was just written and the UI confirms it's set. Verified across repeated calls (no caching involved).
The Table model docs list `dateDependencySettings` as a Table property, and Get base schema is documented as returning the Table model. So I'd expect it to come back on read.
Is `dateDependencySettings` being absent from Get base schema expected, or an oversight? If there's a scope or query param I'm missing, I'd appreciate a pointer. I tried `?include[]=dateDependencySettings` but it returns 422 (invalid value), so it doesn't seem include-gated.
Thanks!