Code Examples

Relevant source files

The following files were used as context for generating this wiki page:

• examples/collection.rs
• examples/database.rs
• examples/query_search.rs
• examples/index_example.rs
• examples/authentication.rs
• src/authentication.rs
• src/database.rs

Introduction

The Code Examples within the Milvus Rust SDK provide developers with practical implementation patterns for interacting with a Milvus vector database. These examples cover the full lifecycle of data management, including client initialization, authentication, database and collection administration, indexing, and high-performance vector searching.

The examples serve as both a functional verification tool and a reference architecture for building Rust applications that require low-latency vector retrieval. They demonstrate the use of asynchronous programming via tokio and the SDK's type-safe builders for schemas and options. Sources: README.md, examples/collection.rs, examples/query_search.rs

Client Connection and Authentication

The foundational step for any operation is establishing a secure connection to the Milvus server. The SDK supports basic authentication and provides a ClientBuilder for configuring credentials and connection parameters.

Connection Workflow

The following diagram illustrates the sequence for establishing an authenticated client session.

Authentication Management

Developers can programmatically manage users, roles, and privilege groups. User passwords are Base64 encoded before transmission to the server. Sources: src/authentication.rs:69-70, examples/authentication.rs:24-30

Operation	Function	Description
Create User	create_user	Registers a new username and password.
Role Assignment	grant_role	Links a specific role to a user account.
Privilege Control	grant_privilege_v2	Assigns specific permissions (e.g., "Load") to a role.
Group Management	create_privilege_group	Groups related privileges for easier administration.

Sources: src/authentication.rs:68, src/authentication.rs:222, src/authentication.rs:596, examples/authentication.rs:88

Collection and Schema Management

Collections in Milvus require a defined CollectionSchema, which is constructed using the CollectionSchemaBuilder. This includes defining primary keys, vector dimensions, and optional dynamic fields.

Data Insertion and Synchronization

Entities are inserted using FieldColumn structures. To ensure data visibility for queries, the flush operation is used to synchronize the write-ahead log with the data storage. Sources: examples/collection.rs:54-58, examples/query_search.rs:136-138

Indexing and Vector Search

Indexing is critical for performance. The SDK allows defining IndexParams such as IndexType (e.g., IvfFlat) and MetricType (e.g., L2).

Indexing Flow

1. Creation: Call create_index on a specific vector field.
2. Loading: Use load_collection to move the indexed data into Query Nodes for searching.
3. Execution: Perform search or query operations.

Searching vs. Querying

The SDK distinguishes between similarity-based searching and expression-based querying.

• Search: Uses a query vector to find the nearest neighbors based on a metric. Sources: examples/query_search.rs:65-81
• Query: Uses a boolean expression (e.g., "age > 20") to filter entities. Sources: examples/query_search.rs:37-43
• Get: Retrieves specific entities by their primary IDs. Sources: examples/query_search.rs:94-106

Database Administration

The SDK supports multi-tenancy through database management. A client can switch its active context using using_database.

Feature	Method	Key Options
Database Creation	create_database	replica_number, diskquota_mb, max_collections
Context Switching	using_database	Automatically clears schema cache and updates interceptors.
Resource Allocation	alter_database_properties	Modifies existing DB constraints.

Sources: src/database.rs:77-130, src/database.rs:414-437

Conclusion

These code examples demonstrate the SDK's capability to handle complex vector database workflows with Rust's safety and concurrency features. By following the patterns for schema definition, authenticated connection, and indexed searching, developers can build robust applications integrated with Milvus. Sources: examples/collection.rs, examples/authentication.rs, tests/aggressivehpctesting/aggressivetesting.md