Loading…
Loading…
The generated mapping is deterministic and full-field, so the common case needs no manual work. Field-level overrides are for the exceptions — the one field in…
The generated mapping is deterministic and full-field, so the common case needs no manual work. Field-level overrides are for the exceptions — the one field in one operation that needs to behave differently from the default. You author them in the UI; there is no mediation script to write.
A field override is the most specific tier of the conversion-option hierarchy:
System default → per-Service → per-Operation → per-Field
↑ overrides here
Per-field overrides are exposed as chips on the leaf fields of an operation's mapping. They cover the leaf-level conversion options:
xsi:type discriminator (and the discriminator field name)pattern.The body-level options act on the whole response body, so there is no single leaf to attach them to. Set these at the system or operation tier instead:
See the conversion options reference for each option's available tiers.
For most options a per-field override is a single response-side value. Two options are bidirectional — they carry separate request-side and response-side overrides per field, and both are consulted at runtime:
Per-field validation regexes are likewise direction-aware (separate request and response patterns). If you need to split a single-value option by direction for a specific leaf, set the distinction at the operation tier rather than the field tier.
epoch-millis, or one string field to snake_case).Because an override creates a new revision, you can always roll back to a known-good version.
A field override changes both:
pattern; an integer-range-bounds change emits/omits minimum/maximum).So the published contract always matches the runtime behavior.