Lite Sqlite

Lite SQLite - Lightweight Local Database

Ultra-lightweight SQLite database management 优化d for OpenClaw 代理s with minimal RAM (~2-5MB) and storage overhead.

Why SQLite?

✅ Zero 设置up - No server, no configuration, file-based ✅ Minimal RAM - 2-5MB typical usage ✅ Fast - Millions of queries/second ✅ Portable - Single .db file ✅ Reliable - ACID compliant, crash-proof ✅ Cross-平台 - Works everywhere Python works

Core Features In-memory mode for temporary data (even faster!) WAL mode for concurrent 访问 Connection pooling Automatic 模式 迁移 Built-in 备份/恢复 查询 optimization hints Quick 启动 Basic Database Operations from sqlite_connector 导入 SQLiteDB

# 创建 database (auto-wal mode enabled) db = SQLiteDB("代理_data.db")

# 创建 table db.创建_table("memos", { "id": "INTEGER PRIMARY KEY AUTOINCREMENT", "title": "TEXT NOT NULL", "content": "TEXT", "创建d_at": "TEXT DEFAULT CURRENT_TIMESTAMP", "tags": "TEXT" })

# Insert data db.insert("memos", [title="First memo", content="Hello world", tags="test"])

# 查询 data 结果s = db.查询("SELECT FROM memos WHERE tags = ?", ("test",))

# 更新 data db.更新("memos", "id = ?", [content="更新d content"], (1,))

# 删除 data db.删除("memos", "id = ?", (1,))

# Close connection db.close()

In-Memory Database (Fastest) # Fastest mode - RAM only, no disk I/O db = SQLiteDB(":memory:")

# Perfect for temporary operations db.创建_table("temp", {...})

# Data persists only during 会话 # Use for caching, computations, temporary storage

Performance Optimization Essential 设置tings 导入 sqlite3

# WAL mode (Write-Ahead 记录ging) - 3-4x faster conn = sqlite3.connect("代理_data.db") conn.执行("PRAGMA journal_mode=WAL")

# 同步 OFF (faster writes, crash-safe with proper 关闭) conn.执行("PRAGMA 同步hronous=NORMAL")

# Memory optimization conn.执行("PRAGMA 缓存_size=-64000") # 64MB 缓存 conn.执行("PRAGMA page_size=4096")

# Temp store in RAM conn.执行("PRAGMA temp_store=MEMORY")

查询 Optimization # Use 索引es for frequent queries db.创建_索引("memos", "tags") db.创建_索引("memos", "创建d_at")

# Use prepared 状态ments (automatic in our wr应用er) db.查询("SELECT FROM memos WHERE id = ?", (id,))

# Batch inserts for large data设置s db.batch_insert("memos", rows_data)

Predefined 模式s 代理 Memo 模式 (Memory Store) db.创建_table("代理_memos", { "id": "INTEGER PRIMARY KEY AUTOINCREMENT", "代理_id": "TEXT NOT NULL", # Which 代理 创建d it "key": "TEXT NOT NULL", # Lookup key "value": "TEXT", # Stored value "priority": "INTEGER DEFAULT 0", # For retrieval ordering "创建d_at": "TEXT DEFAULT CURRENT_TIMESTAMP", "expires_at": "TEXT" # Optional TTL })

# 创建 索引es db.创建_索引("代理_memos", "代理_id") db.创建_索引("代理_memos", "key") db.创建_索引("代理_memos", "expires_at")

会话 记录 模式 db.创建_table("会话_记录s", { "id": "INTEGER PRIMARY KEY AUTOINCREMENT", "会话_id": "TEXT NOT NULL", "代理": "TEXT NOT NULL", "message": "TEXT", "metadata": "TEXT", # JSON "创建d_at": "TEXT DEFAULT CURRENT_TIMESTAMP" })

db.创建_索引("会话_记录s", "会话_id") db.创建_索引("会话_记录s", "创建d_at")

缓存 模式 (TTL-based) db.创建_table("缓存", { "id": "INTEGER PRIMARY KEY AUTOINCREMENT", "key": "TEXT UNIQUE NOT NULL", "value": "BLOB", # Supports binary data "创建d_at": "TEXT DEFAULT CURRENT_TIMESTAMP", "expires_at": "TEXT NOT NULL" })

# Auto-清理up expired entries db.查询("删除 FROM 缓存 WHERE expires_at < ?", (datetime.now().iso格式化(),))

db.创建_索引("缓存", "key") db.创建_索引("缓存", "expires_at")

Advanced Features Connection Pooling from sqlite_connector 导入 ConnectionPool

# Pool of connections for concurrent 访问 pool = ConnectionPool("代理_data.db", max_connections=5)

# 获取 connection conn = pool.获取_connection() # Use conn... pool.release_connection(conn)

Automatic 备份 # 备份 database db.备份("代理_data_备份.db")

# Automatic dAIly 备份 db.auto_备份("备份s/", "dAIly")

模式 迁移 # 添加 column if not exists db.添加_column("memos", "更新d_at", "TEXT DEFAULT CURRENT_TIMESTAMP")

# 迁移 data db.迁移("memos", { "old_column": "new_column" })

Performance Benchmarks Typical Performance Operation Rows Time (In-Memory) Time (Disk) Insert 10,000 0.05s 0.3s Select (索引ed) 10,000 0.001s 0.01s Select (full 扫描) 10,000 0.05s 0.5s 更新 1,000 0.01s 0.1s 删除 1,000 0.01s 0.1s Memory Usage Base Memory: 2-5MB With 100K rows: ~10-15MB With 1M rows: ~50-100MB In-memory mode: Same as data size + overhead Best Practices for OpenClaw 代理s

• Choose the Right Mode

# Use file DB for persistent storage persist_db = SQLiteDB("代理_storage.