Node Driver and Mongoose (Optional)

Introduction

Node.js talks to MongoDB through the official mongodb driver (low level) or Mongoose (schema, middleware, population). This optional chapter shows both with a small blog API and links to Node database basics without repeating npm fundamentals.

Prerequisites

Install

Official driver:

bash

npm install mongodb

Mongoose (ODM):

bash

npm install mongoose

Official Driver: MongoClient

db.mjs:

javascript

Express route:

javascript

Code explanation:

ObjectId.isValid before constructing ObjectId — avoids exceptions and injection-style bugs — Security
Reuse one MongoClient per process (module singleton), not per request

Mongoose: Schema and Model

models/post.js:

javascript

Connect once at startup:

javascript

await mongoose.connect(process.env.MONGODB_URI ?? "mongodb://127.0.0.1:27017/hello_demo");

Create and query:

javascript

const post = await Post.create({
  title: "Mongoose Intro",
  slug: "mongoose-intro",
  published: true,
  tags: ["nodejs", "mongodb"],
});
 
const list = await Post.find({ published: true })
  .sort({ createdAt: -1 })
  .limit(20)
  .select("title slug createdAt")
  .lean();

.lean() returns plain objects—faster read-only APIs.

Population (References)

models/comment.js:

javascript

const commentSchema = new mongoose.Schema({
  postId: { type: mongoose.Schema.Types.ObjectId, ref: "Post", required: true },
  author: String,
  body: String,
  createdAt: { type: Date, default: Date.now },
});
 
export const Comment = mongoose.model("Comment", commentSchema);

javascript

const comments = await Comment.find({ postId: post._id })
  .sort({ createdAt: 1 })
  .populate("postId", "title slug");

Population runs extra queries—at scale prefer aggregation $lookup — chapter 13.

Middleware Example

javascript

postSchema.pre("save", function (next) {
  if (this.isModified("slug")) {
    this.slug = this.slug.toLowerCase().trim();
  }
  next();
});

Hooks replace some app-layer validation; DB validator still helps — Schema Validation.

Driver vs Mongoose

Layer	Best for
`mongodb` driver	Full control, microservices, aggregation-heavy pipelines
Mongoose	CRUD apps, schema enforcement, hooks, quick REST APIs

Tip

Team choice

Pick Mongoose when the team wants schema docs in code. Pick the native driver when you mirror PyMongo-style dict queries or need minimal abstraction.

FAQ

TypeScript?

Use @types/mongoose or Mongoose 7+ built-in types; driver ships TypeScript definitions.

Next.js API routes?

Connect in a shared module; avoid connecting per request in serverless—use connection pooling patterns or Atlas serverless limits.

Validation vs MongoDB validator?

Mongoose validates before send; MongoDB validator is server-side backup — chapter 14.

Fastify instead of Express?

Same driver/Mongoose—swap HTTP layer only.

Compare Python track?

PyMongo Basics and FastAPI integration cover the same CRUD patterns in Python.