Interface CachedRunViewData

Cached aggregate results, if aggregates were requested

The maximum __mj_UpdatedAt timestamp from the results

Set when the slot was written with CacheWriteOptions.ProviderInternalScaffolding. Persisted so that in-place slot maintenance (storeCachedResults, driven by BaseEntity save/delete events) can carry the exemption FORWARD. Without it, the first save event touching a scaffolding slot would freeze it and break the owner that mutates those rows.

The cached result rows.

⚠️ These rows are SHARED: under a reference-sharing storage provider (see ILocalStorageProvider.SharesReferences) they are the same objects held by every reader of this slot, and they are deep-frozen at write time. Mutating one corrupts process-wide state and throws a TypeError. Build a new array / new rows instead.

Deliberately typed mutable rather than readonly: the runtime freeze is the enforcement, and a readonly marker here would be a compile break for existing downstream code that reads cache entries — without adding protection the freeze does not already provide.

Hash of the entity's field names (in sequence order) at the time the cache entry was written. Used to detect schema changes (e.g., new columns added via migration + CodeGen) that would make the cached data structurally stale even though maxUpdatedAt and rowCount haven't changed. Backward-compatible: entries without this field are served normally (no regression).

The newest __mj_UpdatedAt (ISO string) across the entity's metadata row and its field rows, written alongside schemaHash. When two processes sharing a cache disagree on the hash, this says which of them holds the newer metadata, so the older one stops treating the newer one's slots as stale. Absent on entries from writers that predate it or whose metadata carries no timestamps.

Total row count from the database — may differ from results.length for paginated queries