Skip to content

Additional text search algorithm - ngram

The ngram text search algorithm is useful for searching text for a specific string of characters in a field of a collection. This feature can be used to find exact sub-string matches, which provides an alternative to parsing text from languages other than the list of European languages already supported by MongoDB Community’s full text search engine. It may also turn out to be more convenient when working with the text where symbols like dash(‘-‘), underscore(‘_’), or slash(“/”) are not token delimiters.

Unlike MongoDB full text search engine, ngram search algorithm uses only the following token delimiter characters that do not count as word characters in human languages:

  • Horizontal tab
  • Vertical tab
  • Line feed
  • Carriage return
  • Space

The ngram text search is slower than MongoDB full text search.


To use ngram, create a text index on a collection setting the default_language parameter to ngram:

> db.collection.createIndex({name:"text"}, {default_language: "ngram"})

ngram search algorithm treats special characters like individual terms. Therefore, you don’t have to enclose the search string in escaped double quotes (\\") to query the text index. For example, to search for documents that contain the date 2021-02-12, specify the following:

> db.collection.find({ $text: { $search: "2021-02-12" } })

However, both ngram and MongoDB full text search engine treat words with the hyphen-minus - sign in front of them as negated (e.g. “-coffee”) and exclude such words from the search results.

Get expert help

If you need assistance, visit the community forum for comprehensive and free database knowledge, or contact our Percona Database Experts for professional support and services.