The Mental Model

Here's the foundation everything else rests on. If you hold these three words in your head — document, collection, _id — the rest of MongoDB falls into place.

Coming from SQL (or never having touched it)

Here's the translation table I keep in my head:

SQL world	MongoDB world	If you think in JavaScript
Database	Database	—
Table	Collection	an array
Row	Document	an object {}
Column	Field	an object key
Primary key	_id	a unique id
JOIN	$lookup / embedding	nested objects

A document is just a JSON-like object:

{
  "_id": ObjectId("65f0c9a2b3..."),
  "name": "Avina",
  "roles": ["admin", "editor"],
  "profile": { "city": "NYC" },
  "createdAt": ISODate("2026-06-11T10:00:00Z")
}

Notice three things that make it feel familiar if you've worked with JavaScript:

1. Arrays are first-class (roles) — no separate join table.
2. Nested objects are first-class (profile) — the shape mirrors what your UI consumes.
3. No rigid schema — two documents in the same collection don't have to look identical (though in practice you keep them consistent).

The _id field is special

Every document gets a unique _id. If you don't provide one, MongoDB generates a 12-byte ObjectId — and it's cleverer than a random UUID:

Two consequences worth remembering:

• It's roughly time-sortable — sorting by _id ≈ sorting by creation time. You get a free "created at."
• It's generated client-side by the driver, so you know a document's id before it even hits the database.

The size rule

A single document maxes out at 16 MB. That isn't a limitation to fight — it's a design nudge. It stops you from stuffing an unbounded, ever-growing array into one document (which would make every read heavier). When something wants to grow forever, that's your signal to model it differently — which is exactly what the Data Modeling page is about.

What I want you to take away

A MongoDB database holds collections, a collection holds documents, and every document is a JSON-like object with a unique _id. That's it. Hold onto that and the next page — how these documents are actually stored on disk — will make complete sense.

👉 Next: How It Stores Data