Atlas Search Indexes

Overview

In this guide, you can learn how to programmatically manage your Atlas Search and Atlas Vector Search indexes by using the Ruby driver.

The Atlas Search feature enables you to perform full-text searches on collections hosted on MongoDB Atlas. To learn more about Atlas Search, see the Atlas Search Overview.

Atlas Vector Search enables you to perform semantic searches on vector embeddings stored in MongoDB Atlas. To learn more about Atlas Vector Search, see the Atlas Vector Search Overview.

You can call the following methods to manage you Atlas Search indexes:

• search_indexes#create_one

• search_indexes#create_many

• search_indexes#update_one

• search_indexes#drop_one

The following sections provide code examples that demonstrate how to use each of the preceding commands.

Create a Search Index

To create a single Atlas Search or Vector Search index, use the search_indexes#create_one method. To create multiple indexes, use the search_indexes#create_many method. Both methods return immediately, while the indexes are asynchronously created in the background.

The following code example shows how to create an Atlas Search index by providing an index definition and an optional name for the index:

# Creates indexes on all dynamically indexable fields with a default index name
collection.search_indexes.create_one(
  { mappings: { dynamic: true } } 
)
# Creates an index on the specified field with the specified index name
index_definition = {
  mappings: {
    dynamic: false,  
    fields: { 
      fullplot: { type: 'string' }
    }
  }
}
collection.search_indexes.create_one(index_definition, name: 'mySearchIndex')

You can use search_indexes#create_many to create multiple Atlas Search or Vector Search indexes by providing an array of index specifications. Each index specification should include the following components:

• definition parameter: Defines the index

• name parameter: Specifies the index name

• type parameter: Specifies the type of index ('search' or 'vectorSearch')

The following code example shows how to create Atlas Search and Vector Search indexes in one call:

index_spec_1 = {
  name: 'searchIndex_plot',
  type: 'search',
  definition: {
    mappings: {
      dynamic: false,  
      fields: { 
        plot: { type: 'string' }
      }
    }
  }
}
index_spec_2 = {
  name: 'vsIndex_plot_embedding',
  type: 'vectorSearch',
  definition: {
    fields: [
      {
        type: "vector",
        path: "plot_embedding",
        numDimensions: 1536,
        similarity: "dotProduct"
      }
    ]
  }
}
collection.search_indexes.create_many([index_spec_1, index_spec_2])

For longer index definitions, it is helpful to define the index definitions outside of the method call. To learn more about the syntax of index definitions, see the Review Atlas Search Index Syntax guide in the Atlas manual.

Update a Search Index

To update an Atlas Search or Vector Search index, use the search_indexes#update_one method.

To update an index, you must provide a new index definition. You must specify the index you want to update by using either the name or id of the index. The following code shows how to update an Atlas Search index:

updated_definition = {
  mappings: {
    dynamic: false,  
    fields: { fullplot: { type: 'string' } }
    }
}
# Specifies the index to update by using the index name
collection.search_indexes.update_one(updated_definition, name: 'searchIndex_plot')
# Specifies the index to update by using the index id
collection.search_indexes.update_one(updated_definition, id: <index id>)

Delete a Search Index

To delete an Atlas Search or Vector Search index, use the search_indexes#drop_one method.

To delete an index, you must provide the id or name of the index. The following code shows how to delete a search index from a collection:

# Specifies the index to delete by using the index name
collection.search_indexes.drop_one(name: 'searchIndex_plot')
# Specifies the index to delete by using the index id
collection.search_indexes.drop_one(id: <index id>)

List Search Indexes

You can use the search_indexes object to list the entire index specification of each Atlas Search and Vector Search index on a collection:

puts collection.search_indexes.collect(&:to_json)

To list individual fields in the index specification for each index, iterate over the search_indexes object:

collection.search_indexes.each do |index_spec|
  p index_spec['id']
  p index_spec['name']
  p index_spec['status']
  p index_spec['queryable']
  p index_spec['latestDefinition']
end

Additional Information

To learn more about MongoDB Atlas Search, see the Atlas Search documentation.