How to Use Autocomplete in Elasticsearch: Syntax and Examples

To use autocomplete in Elasticsearch, you typically use the completion suggester which provides fast prefix matching for search-as-you-type. Define a completion field in your mapping, index your data with that field, then query using the _search endpoint with a suggest section specifying the prefix.

Syntax

The autocomplete feature in Elasticsearch uses the completion field type in the index mapping. You define a field with type: completion to enable fast prefix suggestions. When querying, you use the _search API with a suggest section that specifies the prefix to autocomplete.

Key parts:

• Mapping: Define a field with type: completion.
• Indexing: Add data to the completion field.
• Query: Use suggest with prefix and completion type.

json

{
  "mappings": {
    "properties": {
      "title": {
        "type": "completion"
      }
    }
  }
}


// Query example
{
  "suggest": {
    "song-suggest": {
      "prefix": "nir",
      "completion": {
        "field": "title"
      }
    }
  }
}

Example

This example shows how to create an index with a completion field, index some documents, and query for autocomplete suggestions with prefix "nir".

json

PUT /music
{
  "mappings": {
    "properties": {
      "title": {
        "type": "completion"
      }
    }
  }
}

POST /music/_doc/1
{
  "title": "Nirvana"
}

POST /music/_doc/2
{
  "title": "Nirvana Live"
}

POST /music/_doc/3
{
  "title": "Nina Simone"
}

POST /music/_search
{
  "suggest": {
    "song-suggest": {
      "prefix": "nir",
      "completion": {
        "field": "title"
      }
    }
  }
}

Output

{
  "suggest": {
    "song-suggest": [
      {
        "text": "nir",
        "offset": 0,
        "length": 3,
        "options": [
          {
            "text": "Nirvana",
            "_index": "music",
            "_id": "1",
            "score": 1.0
          },
          {
            "text": "Nirvana Live",
            "_index": "music",
            "_id": "2",
            "score": 1.0
          }
        ]
      }
    ]
  }
}

Common Pitfalls

Common mistakes when using autocomplete in Elasticsearch include:

• Not defining the field as completion type in the mapping, which disables fast prefix suggestions.
• Indexing data without the completion field populated, so suggestions return empty.
• Using the match query instead of suggest for autocomplete, which is slower and less accurate.
• Not refreshing the index after indexing documents, causing suggestions to miss recent data.

json

POST /music/_search
{
  "query": {
    "match": {
      "title": "nir"
    }
  }
}

POST /music/_search
{
  "suggest": {
    "song-suggest": {
      "prefix": "nir",
      "completion": {
        "field": "title"
      }
    }
  }
}

Quick Reference

Summary tips for Elasticsearch autocomplete:

• Use completion field type in mapping.
• Index data with the completion field populated.
• Query with suggest and prefix for fast autocomplete.
• Refresh index after indexing new data for up-to-date suggestions.

Key Takeaways

Define a field with type completion in your Elasticsearch mapping for autocomplete.

Index your data with the completion field populated to enable suggestions.

Use the suggest API with prefix and completion type for fast autocomplete queries.

Avoid using match queries for autocomplete as they are slower and less precise.

Refresh your index after adding documents to see updated autocomplete results.