Highlighting¶

On this page

Definition
Example

The Atlas Search highlight option adds fields to the result set that display search terms in their original context. You can use it in conjunction with $search operators to display search terms as they appear in the returned documents, along with the adjacent text content (if any). highlight results are returned as part of the $meta field.

Definition¶

highlight has the following syntax:

copy

{
  $search: {
    "index": <index name>, // optional, defaults to "default"
    "<operator>": { // may be ``search``, ``term``, ``compound``, or ``span``
      "query": "<search-string>",
      "path": "<field-to-search>"
    },
    "highlight": { "path": "<field-to-search>" }
  }
},
{
  $project: {
    "highlights": { "$meta": "searchHighlights" } }
  }
}

The path field may contain:

A string
An array of strings
A multi analyzer specification
An array containing a combination of strings and multi analyzer specifications

See Path Construction for more information.

The "$meta": "searchHighlights" field contains the highlighted results. That field isn’t part of the original document, so it is necessary to use a $project pipeline stage to add it to the query output.

Example¶

The following example uses a collection called fruit that contains the following documents:

copy

{
  "_id" : 1,
  "type" : "apple",
  "description" : "Apples come in several varieties, including Fuji, Granny Smith, and Honeycrisp."
},
{
  "_id" : 2,
  "type" : "banana",
  "description" : "Bananas are usually sold in bunches of five or six."
}

Note

One useful aspect of highlighting is that it reveals the original text returned by the search query, which may not be exactly the same as the search term. For example, if you use a language-specific analyzer, your text searches return all the stemmed variations of your search terms.

The fruit collection has an index definition that uses the english analyzer and dynamic field mappings.

copy

{
  "analyzer": "lucene.english",
  "searchAnalyzer": "lucene.english",
  "mappings": {
    "dynamic": true
  }
}

The following query searches for variety and bunch in the description field of the fruit collection, with the highlight option enabled.

The $project pipeline stage restricts the output to the description field and adds a new field called highlights, which contains highlighting information.

copy

db.fruit.aggregate([
  {
    $search: {
      "search": {
        "path": "description",
        "query": ["variety", "bunch"]
      },
      "highlight": {
        "path": "description"
      }
    }
  },
  {
    $project: {
      "description": 1,
      "_id": 0,
      "highlights": { "$meta": "searchHighlights" }
    }
  }
])

The query returns the following results:

{
  "description": "Bananas are usually sold in bunches of five or six.",
  "highlights" : [
    {
      "path" : "description",
      "texts" : [
        {
          "value" : "Bananas are usually sold in ",
          "type" : "text"
        },
        {
          "value" : "bunches",
          "type" : "hit"
        },
        {
          "value" : " of five or six.",
          "type" : "text"
        }
      ],
      "score" : 1.2841906547546387
    }
  ]
}
{
  "description" : "Apples come in several varieties, including Fuji, Granny Smith, and Honeycrisp.",
  "highlights" : [
    {
      "path" : "description",
      "texts" : [
        {
          "value" : "Apples come in several ",
          "type" : "text"
        },
        {
          "value" : "varieties",
          "type" : "hit"
        },
        {
          "value" : ", including Fuji, Granny Smith, and Honeycrisp.",
          "type" : "text"
        }
      ],
      "score" : 1.2178014516830444
    }
  ]
}

The highlights field is an array containing the following output fields:

Field	Type	Description
`path`	string	Document field which returned a match.
`texts`	array of objects	Each search match returns one or more objects, containing the matching text and the surrounding text (if any).
`texts.value`	string	Text from the field which returned a match.
`texts.type`	string	Returned value is either `hit` or `text`. Results of type `hit` contain the string which returned a match. Results of type `text` contain the text content adjacent to the matching string.
`score`	float	The score assigned to the `highlights` object. The `highlights` score is a measure of the relevance of the `highlights` object to the query. If multiple `highlights` objects are returned, the most relevant `highlights` object has the highest score.

The search term bunch returns a match on the document with _id: 2, because the description field contains the word bunches. The search term variety returns a match on the document with _id: 2, because the description field contains the word varieties. See stemming.