Modeling Many-to-Many Relationships in MongoDB: From SQL to NoSQL

In relational databases a many‑to‑many relationship between articles and tags is usually represented by three tables: Article (Id, Title, …), Tag (Id, Name) and a junction table Relation (ArticleId, TagId). By joining these tables you can query any combination of articles and tags.

MongoDB does not support joins, so the schema must be adapted. A common pattern is to store an array of tag identifiers inside each article document and optionally keep a reverse reference in the tag document.

{
  "id": "505cad",
  "title": "test title",
  "tags": ["tag1", "tag2"]
}

{
  "tag": "tag1",
  "article": ["article1", "article2"],
  "size": 2
}

Example queries:

List all tags: db.article.distinct("tags") → ["tag1", "tag2", "tag3"]

List all articles with their tags: db.article.find({}) returns each document’s _id, title and tags array.

Find articles that contain a specific tag (e.g., "tag2"): db.article.find({tags: {$in: ["tag2"]}}) returns the matching article documents.