Redis
redis-vl-dotnet Documentation redis-vl-dotnet

redis-vl-dotnet

Index Lifecycle

SearchIndex owns the index lifecycle once you have a SearchSchema and a Redis IDatabase.

Constructing an index handle

var schema = SearchSchema.FromYamlFile(schemaPath);
var index = new SearchIndex(database, schema);

This only creates the client-side handle. Redis index creation happens when you call Create or CreateAsync.

Creating an index

The basic flow is:

await index.CreateAsync(new CreateIndexOptions(
    overwrite: true,
    dropExistingDocuments: true));

CreateIndexOptions supports three behaviors:

  • skipIfExists: return false instead of failing when the index already exists

  • overwrite: drop and recreate the index when it already exists

  • dropExistingDocuments: when overwrite drops an existing index, also delete matching Redis documents

skipIfExists and overwrite are mutually exclusive.

Inspecting and rediscovering indexes

Use these APIs after creation:

  • ExistsAsync() checks whether the current schema index exists

  • InfoAsync() returns SearchIndexInfo, a key-value view over FT.INFO

  • SearchIndex.ListAsync(database) returns all index names as SearchIndexListItem

  • SearchIndex.FromExistingAsync(database, indexName) recreates a SearchIndex from Redis metadata instead of a local schema definition

The primary example uses both ListAsync and FromExistingAsync so you can verify what Redis created before running document operations.

Clearing versus dropping

Use the right cleanup method for the goal:

  • ClearAsync() deletes indexed documents by prefix but leaves the index in place

  • DropAsync() removes the index itself

  • DropAsync(deleteDocuments: true) removes both the index and the underlying documents

ClearAsync() is useful in tests or repeatable examples when you want to keep the index definition but wipe data. DropAsync() is the final teardown step when you no longer need the index metadata at all.

Storage-specific document lifecycle

After the index exists, document operations follow the schema storage type:

  • JSON: LoadJsonAsync, FetchJsonByIdAsync, UpdateJsonByIdAsync, DeleteJsonByIdAsync

  • HASH: LoadHashAsync, FetchHashByIdAsync, UpdateHashByIdAsync, DeleteHashByIdAsync

The JSON example demonstrates create, list, rediscover, clear, and drop flows. The HASH vector example shows the same lifecycle with hash-backed documents.

Example reference

Primary example:

  • /examples/JsonStorageExample

Supporting HASH example:

  • /examples/VectorSearchExample

Related pages