Embedding vs Referencing

Data Modeling: Embedding vs Referencing

One of the most important MongoDB design decisions is whether to embed related data inside a document or store it separately and reference it by _id. The right choice depends on access patterns, document size, and how often the data changes.

Embedding (Denormalisation)

Embed when the sub-data is always accessed with the parent, changes together with it, and the total document stays under 16 MB.

// Embed address directly in the student document
{
  name: "Aman Verma",
  address: {
    street: "45 Lake View",
    city: "Pune",
    pincode: "411001"
  }
}

Referencing (Normalisation)

Reference when the related data is large, shared by many documents, or updated independently.

// Store courseId as a reference
{
  name: "Priya Singh",
  courseId: "fs101"   // ObjectId or string key
}

// Courses live in their own collection
{ _id: "fs101", title: "Full Stack", instructor: "Rohan" }