Full-Text Search in MongoDB

Full-Text Search in MongoDB

MongoDB provides built-in full-text search capabilities through text indexes. A text index tokenizes and stems string content, allowing you to search for words and phrases across one or more fields. This is faster and more capable than using $regex for keyword searches.

Creating a Text Index

A collection can have at most one text index, but it can cover multiple fields:

// Text index on a single field
db.articles.createIndex({ content: "text" })

// Text index on multiple fields
db.articles.createIndex({
  title: "text",
  content: "text",
  tags: "text"
})

// With weights to boost certain fields
db.articles.createIndex(
  { title: "text", content: "text" },
  { weights: { title: 10, content: 1 } }
)

Basic Text Search

Use the $text operator with $search to find documents containing specific words:

// Search for documents containing "mongodb" or "tutorial"
db.articles.find({
  $text: { $search: "mongodb tutorial" }
})

// Exact phrase search (use escaped quotes)
db.articles.find({
  $text: { $search: '"aggregation pipeline"' }
})

// Exclude a term with the minus prefix
db.articles.find({
  $text: { $search: "mongodb -deprecated" }
})

By default, multiple words are treated as an OR -- the document must contain at least one of the terms.

Text Score and Sorting

MongoDB assigns a relevance score to each matching document. Use $meta: "textScore" to access and sort by this score:

// Return results sorted by relevance
db.articles.find(
  { $text: { $search: "mongodb performance" } },
  { score: { $meta: "textScore" } }
).sort({ score: { $meta: "textScore" } })

// Result includes a "score" field:
// { title: "MongoDB Performance Tips", score: 3.5, ... }
// { title: "Database Performance", score: 1.2, ... }

Language Support

Text indexes support language-specific stemming and stop words. You can set the language at the index or query level:

// Create a text index with Spanish language
db.articles.createIndex(
  { content: "text" },
  { default_language: "spanish" }
)

// Search with a specific language
db.articles.find({
  $text: {
    $search: "base de datos",
    $language: "spanish"
  }
})

// Disable stemming with "none"
db.articles.find({
  $text: {
    $search: "running",
    $language: "none"  // matches "running" exactly, not "run"
  }
})

Text Search in Aggregation

Combine text search with the aggregation pipeline for advanced filtering and analysis:

db.articles.aggregate([
  { $match: { $text: { $search: "mongodb" } } },
  { $addFields: { score: { $meta: "textScore" } } },
  { $match: { score: { $gt: 2.0 } } },
  { $sort: { score: -1 } },
  { $limit: 10 },
  { $project: { title: 1, score: 1 } }
])

Key Takeaways

• A collection can have only one text index, but it can span multiple fields.
• Use $text with $search for word and phrase matching.
• Sort by textScore to rank results by relevance.
• For more advanced search needs, consider MongoDB Atlas Search with Lucene-based indexes.