Guide

DictSQLite is a fast persistent dictionary backed by SQLite. Version 2 is implemented with Rust/PyO3 and provides both sync and async APIs.

Quick links

• API reference: Operations for DictSQLite, AsyncDictSQLite, and TableProxy
• Choose modes: Pick pickle/jsonb/json/bytes and lazy/writethrough modes
• Use async: set/get/batch/flush for asyncio workloads
• Read benchmarks: Measurements by size, count, batch, and storage format

Installation

pip install dictsqlite

Minimal example

from dictsqlite import DictSQLite

db = DictSQLite("cache.db")
db["user:1"] = {"name": "Alice", "score": 42}

print(db["user:1"])
db.close()

Async API

import asyncio
from dictsqlite import AsyncDictSQLite

async def main():
    db = AsyncDictSQLite("cache.db")
    await db.set("job:1", {"status": "queued"})
    print(await db.get("job:1"))
    await db.close()

asyncio.run(main())

Storage and persistence modes

• storage_mode: pickle, jsonb, json, bytes
• persist_mode: memory, lazy, writethrough
• table_mode: prefix, separate

Use Safe Pickle and encryption according to your trust boundary. lazy is useful for heavy write throughput, while writethrough is best when each write must be persisted immediately.

Common setups

Use case	Recommended setup
Temporary cache	persist_mode="memory"
Fast persistent cache	persist_mode="lazy", storage_mode="jsonb"
Durability after every write	persist_mode="writethrough"
Binary data	storage_mode="bytes"
Separated tables	table_mode="separate"