How to Use Index Sorting for Performance in Elasticsearch

Use index.sort in your Elasticsearch index settings to pre-sort documents by one or more fields. This helps Elasticsearch quickly return sorted results without extra sorting at query time, improving performance especially for large datasets.

Syntax

The index.sort setting defines how documents are sorted when indexed. It includes:

• fields: An array of fields to sort by, in order of priority.
• order: Sort direction, either asc (ascending) or desc (descending).

This sorting is applied at index time, so queries that sort on these fields run faster.

json

{
  "settings": {
    "index": {
      "sort": {
        "fields": ["field1", "field2"],
        "order": ["asc", "desc"]
      }
    }
  }
}

Example

This example creates an index with index sorting on timestamp ascending and user_id descending. Queries sorting by these fields will be faster.

json

PUT /my-index
{
  "settings": {
    "index": {
      "sort": {
        "fields": ["timestamp", "user_id"],
        "order": ["asc", "desc"]
      }
    }
  },
  "mappings": {
    "properties": {
      "timestamp": { "type": "date" },
      "user_id": { "type": "keyword" },
      "message": { "type": "text" }
    }
  }
}

Output

{"acknowledged":true,"shards_acknowledged":true,"index":"my-index"}

Common Pitfalls

1. Index sorting cannot be changed after index creation. You must define it when creating the index.

2. Sorting on fields not defined in index.sort.fields does not benefit from index sorting. Queries sorting on other fields still require runtime sorting.

3. Using index sorting on high-cardinality fields may increase indexing time and disk usage. Choose fields wisely.

json

PUT /wrong-index
{
  "settings": {
    "index": {
      "sort": {
        "fields": ["nonexistent_field"],
        "order": ["asc"]
      }
    }
  }
}

# Correct approach:
PUT /correct-index
{
  "settings": {
    "index": {
      "sort": {
        "fields": ["timestamp"],
        "order": ["asc"]
      }
    }
  }
}

Output

{"error":{"root_cause":[{"type":"illegal_argument_exception","reason":"Field [nonexistent_field] is not mapped"}],"type":"illegal_argument_exception","reason":"Field [nonexistent_field] is not mapped"}}

Quick Reference

Setting	Description	Notes
index.sort.fields	Fields to sort by at index time	Must be mapped fields
index.sort.order	Sort order for each field	Use 'asc' or 'desc'
Immutable	Index sorting cannot be changed after creation	Plan before indexing
Performance	Speeds up queries sorting on these fields	Best for frequent sort queries
Limitations	Increases indexing cost and disk usage	Avoid high-cardinality fields

Key Takeaways

Define index.sort fields and order when creating the index to improve sorted query speed.

Index sorting only helps queries sorting on the specified fields; other sorts still require runtime sorting.

You cannot change index sorting after index creation, so plan your sorting fields carefully.

Using index sorting may increase indexing time and disk space, especially with many unique values.

Choose fields with moderate cardinality and frequent sort usage for best performance benefits.