InMemoryDocumentStore<T>

Models & Types · AiDotNet.RetrievalAugmentedGeneration.DocumentStores

High-performance in-memory vector document store optimized for rapid prototyping and small-to-medium datasets.

Inheritance: DocumentStoreBase<T> → InMemoryDocumentStore<T>

For Beginners

An in-memory store keeps all your data in RAM (like variables in your program).

Think of it like a super-fast notebook that lives in your computer’s memory:

  • Lightning-fast access (microseconds vs milliseconds)
  • Zero configuration - just create and use
  • Perfect for prototyping and testing
  • Data is lost when the program stops

Best used for:

  • Development and testing
  • Session-based temporary data
  • Caching frequently accessed vectors
  • Small to medium datasets (< 100K documents)

Key characteristics:

  • O(log n) similarity search using HNSW approximate nearest neighbor indexing
  • Thread-safe concurrent access
  • Low memory overhead with efficient storage
  • Simple serialization/deserialization support

How It Works

This implementation provides a pure in-memory vector store using .NET’s high-performance collections, offering extremely fast vector similarity search with minimal configuration. Ideal for development, testing, caching layers, and applications with datasets that fit comfortably in RAM.

Constructors

Constructor Summary
InMemoryDocumentStore(int vectorDimension, string databasePath, string tableName, int hnswMaxConnections, int hnswEfConstruction, int hnswEfSearch) Initializes a new instance of the InMemoryDocumentStore class.

Properties

Property Summary
int DocumentCount { get; } Gets the number of documents currently stored in the in-memory store.
int VectorDimension { get; } Gets the dimensionality of vectors stored in this database.

Methods

Method Summary
void Clear() Removes all documents from the database and resets the vector dimension.