DEV Community

魔眼天王
魔眼天王

Posted on

Core Database Operation Solutions in HarmonyOS

#harmonyos

I. Native API Deep Dive

HarmonyOS provides two native database solutions: ​Relational Database (RDB)​​ and ​Object-Relational Mapping (ORM)​, both built on SQLite.

1. Relational Database (RDB)

​Core API Implementation:

// Database Initialization (TypeScript) import relationalStore from '@ohos.data.relationalStore'; const config = { name: 'user.db', securityLevel: relationalStore.SecurityLevel.S1 }; async function initDB() { const store = await relationalStore.getRdbStore(config); await store.executeSql( 'CREATE TABLE IF NOT EXISTS user (id INTEGER PRIMARY KEY AUTOINCREMENT, name TEXT)' ); } // Batch Insert with Transactions async function batchInsert(users: string[]) { const store = await relationalStore.getRdbStore(config); await store.beginTransaction(); try { users.forEach(name => { store.insert('user', { name }); }); await store.commit(); } catch (e) { await store.rollback(); } }
Enter fullscreen mode Exit fullscreen mode

​Features:

  • ​Advantages:

    • Direct SQL control for complex queries (e.g., multi-table joins)
    • Full ACID transaction support
    • Distributed sync capability (API 18+)

  • ​Limitations:

    • Manual SQL injection prevention
    • Strict single-writer/multi-reader rules for multi-threading
2. Object-Relational Mapping (ORM)

​Code Example:

// Entity Definition @Entity(tableName = 'product') class Product extends OrmObject { @PrimaryKey(autoGenerate = true) private id: number; @Column(String) name: string; } // Database Operations const ormContext = ormCreateDb(); const product = new Product(); product.name = 'Smart Watch'; await ormContext.insert(product); const results = await ormContext.query( ormContext.where(Product).equalTo('name', 'Smart Watch') );
Enter fullscreen mode Exit fullscreen mode

​Advantages:

  • 50% less boilerplate code (no manual SQL)

  • Compile-time type safety

  • Automatic object serialization

​*​

II. Third-Party Libraries Comparison

1. ObjectBox (Community Recommendation)

​Integration:

// package.json "dependencies": { "com.objectbox.objectbox": "1.6.0" }
Enter fullscreen mode Exit fullscreen mode

​Code Example:

// Entity Definition @Entity class User { @Id(autoincrement = true) id: number; name: string; } // Operations const boxStore = MyObjectBox.builder().build(); const userBox = boxStore.boxFor(User.class); userBox.put(new User("Zhang San")); const users = userBox.query().build().find();
Enter fullscreen mode Exit fullscreen mode

​Performance:

  • Insertion: 1,800 entries/sec vs. Native API's 1,200

  • Query Latency: 18ms vs. 25ms

​Use Cases:

  • High-frequency IoT logging

  • Complex object relationships

2. HSQLDB (Lightweight Option)

​Features:

  • Pure Java implementation (no native dependencies)

  • In-memory database support

import { HSQLDatabase } from '@ohos.hsqldb'; const db = new HSQLDatabase(); await db.connect('jdbc:hsqldb:mem:test'); await db.execute('CREATE TABLE session (id INT PRIMARY KEY)');
Enter fullscreen mode Exit fullscreen mode

​Limitations:

  • Limited transaction support (row-level locking only)

  • No distributed deployment

​*​

III. Performance & Scenario Comparison

Aspect Native API ORM Framework ObjectBox
Development ★★☆ (SQL required) ★★★★ (Type-safe) ★★★☆
Speed ★★★★ (Direct) ★★★ (ORM overhead) ★★★★☆
Learning ★★★ (SQL needed) ★★☆ (Annotations) ★★★☆
Complex Query ★★★★ (JOIN support) ★★☆ (Generator) ★★★☆
Distributed ★★★★ (API 18+) ❌ ❌

​*​

IV. Selection Guide

  1. ​Native API:

  • Ideal for complex data models needing distributed sync

  • Use @Transaction for batch operations

  1. ​ORM Framework:

  • Top choice for small/medium projects (TypeScript decorators)

  • Note: Manual Date type conversions

  1. ​Third-Party:

  • ​ObjectBox: IoT local storage

  • ​HSQLDB: Memory-light apps

​Optimizations:

  • Use executeBatch() instead of loops (+3x throughput)

  • Specify columns in queries to avoid full scans

  • Enable WAL mode (PRAGMA journal_mode=WAL) for concurrency

Top comments (0)