File-based Database (JSON)

Before introducing a real database, persist data to disk as JSON. Learn the limitations (concurrency, scale) and when to graduate to SQLite/PostgreSQL.

persistenceJSONfslowdb

Duration

⏱ About 1.5 hours

Level

📊 Intermediate

Prerequisite

🎯 node-05

OUTCOME

Persist a REST API's data to a JSON file (good enough for prototypes)

What you'll learn

1Read/write a JSON file with fs.promises
2Wrap I/O in helper functions
3Recognize race-condition limitations
4Migrate naturally to a real database later

1. A Tiny JSON Store

javascript

import fs from "node:fs/promises";
const FILE = "./data.json";

async function load() {
  try {
    return JSON.parse(await fs.readFile(FILE, "utf8"));
  } catch {
    return { tasks: [], nextId: 1 };
  }
}

async function save(data) {
  await fs.writeFile(FILE, JSON.stringify(data, null, 2));
}

const db = await load();
db.tasks.push({ id: db.nextId++, title: "buy milk" });
await save(db);

2. Integrate with Express

javascript

app.get("/tasks", async (req, res) => {
  const db = await load();
  res.json(db.tasks);
});
app.post("/tasks", async (req, res) => {
  const db = await load();
  const task = { id: db.nextId++, title: req.body.title, done: false };
  db.tasks.push(task);
  await save(db);
  res.status(201).json(task);
});

3. Limitations

Two requests writing at the same time can lose data (race condition)
Whole file is read/rewritten on every operation — slow past a few MB
No query language, no indexes
Fine for: personal tools, prototypes, low-traffic dashboards
Not fine for: any production app with concurrent users

💡

Next module — Databases — graduates from JSON to SQLite (single file but with SQL) and beyond.

Example code / lecture materials

All lecture materials and example code are openly available on GitHub.

View on GitHub ↗