Kinesis DB

Warning

Multiple additions and updates have been made to this project since the initial commits but are only present in the production version of Kinesis API. Please consult the technical documentation to get a better overview of the complete featureset and capabilities.

A high-performance, ACID-compliant embedded database written in Rust with support for multiple storage engines and flexible schema management.

Features

• Multiple Storage Engines

  • In-Memory: Fast, volatile storage for temporary data
  • On-Disk: Persistent storage with ACID guarantees
  • Hybrid: Intelligent caching with persistent backing

• Schema Management

  • Dynamic schema creation and updates
  • Strong type system with constraints
  • Support for required fields, unique constraints, and default values
  • Pattern matching for string fields
  • Min/max value constraints for numeric fields

• Query Interface

  • SQL-like command syntax for familiarity
  • Table operations (CREATE, DROP)
  • Record operations (INSERT, UPDATE, DELETE)
  • Search capabilities across string fields
  • Multiple output formats (Standard, JSON, Table)

• Transaction Support

  • ACID compliance
  • Configurable isolation levels
  • Deadlock detection
  • Write-Ahead Logging (WAL)

REPL Command Reference

Use the built-in help system to learn about available commands:

-- Show all commands
HELP

-- Get details about a specific command
HELP CREATE_TABLE

Common commands:

-- Create a table
CREATE_TABLE users
    name STRING --required
    age INTEGER --min=0 --max=150

-- Insert a record
INSERT INTO users ID 1 SET
    name = "Alice"
    age = 25

-- Query records
GET_RECORDS FROM users
GET_RECORD FROM users 1
SEARCH_RECORDS FROM users MATCH Alice

-- Update a record
UPDATE users ID 1 SET age = 26

-- Delete a record
DELETE FROM users 1

-- Update schema
UPDATE_SCHEMA users --version=2
    active BOOLEAN --default=true

Configuration Options

Storage Engine Selection

• InMemory: Fastest performance, no persistence
• OnDisk: Full persistence with ACID guarantees
• Hybrid: Balanced performance with persistence

Transaction Isolation Levels

• ReadUncommitted: Highest performance, lowest isolation
• ReadCommitted: Prevents dirty reads
• RepeatableRead: Prevents non-repeatable reads
• Serializable: Highest isolation, may impact performance

Buffer Pool Configuration

• Adjust size based on available memory
• Consider workload patterns
• Monitor eviction rates

Performance Considerations

• Choose appropriate storage engine for your use case
• Configure buffer pool size based on available memory
• Use appropriate isolation level for your requirements
• Consider batch operations for bulk data handling
• Index fields that are frequently searched

Error Handling

The database provides detailed error messages for:

• Schema validation failures
• Constraint violations
• Deadlock detection
• Transaction conflicts
• I/O errors

Contributing

Pull requests are welcome! Please ensure your changes include:

• Tests for new functionality
• Documentation updates
• Proper error handling
• Code formatting with cargo fmt
• Pass all checks with cargo test