Orchardcore Query Indexing

OrchardCore Query & Indexing

This skill guides you through OrchardCore's query and indexing systems following project conventions.

Two distinct systems, often confused:

• Queries (OrchardCore.Queries) — named, stored queries executed on demand (SQL, Lucene, Elasticsearch sources). Return raw rows or content items.
• Indexing (OrchardCore.Indexing + providers) — build full-text index profiles that mirror content into Lucene/Elasticsearch/Azure AI Search for searching. Driven by index handlers.

Separate again from YesSql SQL index tables (MapIndex + SchemaBuilder) — those are relational indexes for fast DB queries, covered by the orchardcore-data-migration skill, not here.

Decide what you need

Workflow A: define and run a query

Step 1: Pick a source

Query has Name, Source (Sql/Lucene/Elasticsearch), ReturnContentItems, and source-specific metadata (e.g. a SQL Template).

Step 2: Create via recipe or manager

Recipe step:

{
  "name": "queries",
  "Queries": [
    {
      "Source": "Sql",
      "Name": "RecentPosts",
      "Template": "select DocumentId from ContentItemIndex where ContentType = 'BlogPost'",
      "ReturnContentItems": true
    }
  ]
}

Programmatically:

var query = await _queryManager.NewAsync("Sql", data);
await _queryManager.SaveAsync(query);

Step 3: Execute

var query = await _queryManager.GetQueryAsync("RecentPosts");
var results = await _queryManager.ExecuteQueryAsync(query, new Dictionary<string, object>
{
    ["maxResults"] = 10,
});
foreach (var item in results.Items) { /* JsonObject or ContentItem */ }

SQL templates are Liquid-rendered (parameters become Liquid vars), parsed by SqlParser, then run via Dapper. ReturnContentItems = true loads ContentItems by DocumentId; otherwise rows are returned as JsonObject.

Workflow B: index content for search

Step 1: Create an index profile

A profile binds a provider (Lucene/Elasticsearch/AzureAISearch) + a type (Content) + which content types to index.

{
  "name": "CreateOrUpdateIndexProfile",
  "indexes": [
    {
      "Name": "BlogPostsLucene",
      "IndexName": "blogposts",
      "ProviderName": "Lucene",
      "Type": "Content",
      "Properties": {
        "ContentIndexMetadata": {
          "IndexLatest": false,
          "IndexedContentTypes": [ "BlogPost" ],
          "Culture": "any"
        },
        "LuceneIndexMetadata": { "AnalyzerName": "standard", "StoreSourceData": true }
      }
    }
  ]
}

Content changes raise indexing tasks; ContentIndexingService resolves the content item and calls every part/field index handler to build a DocumentIndex, which the provider stores.

Step 2: Index custom part/field data

For a custom part, implement ContentPartIndexHandler<TPart>:

public class AliasPartIndexHandler : ContentPartIndexHandler<AliasPart>
{
    public override Task BuildIndexAsync(AliasPart part, BuildPartIndexContext context)
    {
        var options = DocumentIndexOptions.Keyword | DocumentIndexOptions.Store;
        foreach (var key in context.Keys)
        {
            context.DocumentIndex.Set(key, part.Alias, options);
        }
        return Task.CompletedTask;
    }
}

Fields use ContentFieldIndexHandler<TField> with BuildFieldIndexContext. context.Keys are the index field names to populate; DocumentIndexOptions flags control storage:

Step 3: Register the handler

services.AddScoped<IContentPartIndexHandler, AliasPartIndexHandler>();
// field:
services.AddScoped<IContentFieldIndexHandler, MyFieldIndexHandler>();

Step 4: Reset / rebuild

Recipe steps ResetIndex (re-sync from last position) and RebuildIndex (drop + rebuild) manage existing indexes. The admin Search/Indexing UI exposes the same.

Quick Reference

Query APIs

Indexing APIs

Register an indexing source

services.AddIndexingSource<TIndexManager, TDocumentIndexManager, TIndexNameProvider>(
    providerName, implementationType);
// e.g. the Lucene module: services.AddLuceneIndexingSource(implementationType);

Search

ISearchService.SearchAsync(IndexProfile index, string term, int start, int size) → SearchResult. Providers: OrchardCore.Search.Lucene, .Elasticsearch, .AzureAI.

Gotchas

• "Index" is overloaded: full-text index profiles (this skill) vs YesSql MapIndex tables (orchardcore-data-migration). Don't mix them.
• Query recipe step is queries; index profile step is CreateOrUpdateIndexProfile. Older lucene-index / ElasticIndexSettings steps are deprecated — prefer CreateOrUpdateIndexProfile.
• IndexLatest: true indexes draft (latest) versions; false indexes published. Match to your search UI.
• SQL query Template is Liquid first, then SQL — parameters are Liquid variables, not ADO placeholders.

References

• references/indexing.md — index profiles, handlers, DocumentIndex, providers, recipe steps
• src/docs/reference/modules/Queries/README.md (repo)
• src/docs/reference/modules/Indexing/README.md (repo)
• src/docs/reference/modules/{Lucene,Elasticsearch,SQLIndexing}/README.md (repo)
• AGENTS.md (repo root) — build commands