Skip to content
Aurora Catalyst

Auto-include FK columns in lists

Auto-generated from the source archive. Do not edit by hand — rerun catalyst-changelog-sync instead.

What changed

Section titled “What changed”
  • For every many-to-one relationship, <mod>.columns.ts now emits a column with accessorKey: '<rel>.name', relation: { association: '<rel>' }, and the standard sortable / searchable / filterable / exportable flags. Position follows the FK’s declaration order in the YAML.
  • The list-component imports getRelationIncludes from @aurora, declares a relationIncludes field, and passes include in both query and constraint of the paginate() call. The list resolver wires the same include for the first-page pre-load so the table renders without a flicker.
  • Search and filter on FK columns work without any extra codegen — the existing buildSearchWhere / buildFilterWhere helpers translate boundedContext.name to Sequelize $boundedContext.name$. Aggregates with no many-to-one FKs regenerate byte-identical.

Why it matters

Section titled “Why it matters”

Before this, the list of an aggregate with FKs showed the local fields but hid the <rel>Id UUIDs — a permissions list did not display the bounded-context each row belonged to, which made it useless as an exploration tool. Devs wrote the FK column by hand and every regen produced a .origin to merge. The codegen now owns it end to end: column emission, resolver include for the first-page pre-load, list-component include for runtime paginate, and dot-path translation in search/filter. The label is hardcoded to the target’s name field; aggregates with composite labels ({code} - {name}) or without name are deferred to a follow-up.

View original proposal