When you change your type registrations (add a new [Queryable] property, register a new type, etc.), LottaDB detects the schema change on startup and automatically rebuilds the Lucene index from table storage.

// v1: only Actor registered
var db = await catalog.GetDatabaseAsync("main", c => c.Store<Actor>());

// v2: added Note -- schema changed, index auto-rebuilt
var db = await catalog.GetDatabaseAsync("main", c =>
{
    c.Store<Actor>();
    c.Store<Note>();  // new! triggers rebuild
});

The schema hash includes an IndexFormatVersion constant so Lucene document format changes (e.g., adding new stored fields) also trigger a rebuild on upgrade.

JsonSchema entities are loaded from Table Storage at startup and included in the schema hash. When you update a JsonSchema at runtime, the built-in On<JsonSchema> handler detects the change and triggers a selective reindex -- only that document type's Lucene entries are rebuilt, not the entire index.

// Add a new queryable property -- triggers reindex of Person documents only
await db.SaveAsync(new JsonSchema
{
    Name = "Person",
    Properties = new()
    {
        new() { Name = "Name", Type = "string" },
        new() { Name = "Age", Type = "integer" },
        new() { Name = "Email", Type = "string" },  // new!
    }
});

Deleting a JsonSchema removes its mapper and Lucene index entries.

You can force a full rebuild at any time:

await db.RebuildSearchIndex();

This deletes the entire Lucene index and re-indexes all entities (typed and dynamic) from Table Storage in a single pass. Does not run On<T> handlers.