Link Search Menu Expand Document Documentation Menu
Documentation

Transform string fields to text/keyword

This guide explains how Migration Assistant automatically handles the deprecated string field type during migration from earlier Elasticsearch versions.

Overview

The string field type was the primary text field type in early versions of Elasticsearch but was deprecated in Elasticsearch 5.x and completely removed in Elasticsearch 6.0. In Elasticsearch 5.x, it remained available in backward compatibility mode only.

When migrating from Elasticsearch 1.x through 5.x to later versions, Migration Assistant automatically converts string field types to their modern equivalents: text for analyzed fields and keyword for non-analyzed fields.

To determine whether an Elasticsearch cluster uses string field types, make a call to your source cluster’s GET /_mapping API. In the migration console, run console clusters curl source_cluster "/_mapping". If you see "type":"string", then this transformation is applicable and these fields will be automatically transformed during migration.

Compatibility

The string to text/keyword field type transformation applies to:

  • Source clusters: Elasticsearch 1.x–5.x
  • Target clusters: Elasticsearch 5.x+ or OpenSearch 1.x+
  • Automatic conversion: No configuration required during metadata migrations.

Automatic conversion logic

Migration Assistant determines whether to convert a string field to text or keyword based on the field’s index property.

Conversion to keyword

Fields are converted to keyword when:

  • index: "not_analyzed"
  • index: "no"
  • index: false

Conversion to text

Fields are converted to text when:

  • index: "analyzed"
  • index is undefined

Property cleanup

During conversion, Migration Assistant also cleans up properties that are incompatible with the target field type.

For keyword fields, the following properties are removed:

  • analyzer
  • search_analyzer
  • position_increment_gap
  • term_vector
  • fielddata

For text fields, the following properties are removed:

  • doc_values
  • null_value (if present)

Additionally, legacy index values are normalized:

  • index: "analyzed" and index: "not_analyzed" are removed (Elasticsearch 5.x default is true).
  • index: "no" becomes index: false.

Migration output

During the migration process, you’ll see this transformation in the output:

Transformations:
   string to text/keyword:
      Convert field type string to text/keyword based on field data mappings

Transformation behavior

Source field type Target field type 
{
  "properties": {
    "status": {
      "type": "string",
      "index": "not_analyzed",
      "doc_values": true
    },
    "category": {
      "type": "string",
      "index": "no"
    }
  }
}
 
{
  "properties": {
    "status": {
      "type": "keyword",
      "doc_values": true
    },
    "category": {
      "type": "keyword",
      "index": false
    }
  }
}
 
{
  "properties": {
    "title": {
      "type": "string",
      "analyzer": "standard",
      "fielddata": true
    },
    "description": {
      "type": "string",
      "index": "analyzed",
      "term_vector": "with_positions"
    }
  }
}
 
{
  "properties": {
    "title": {
      "type": "text",
      "analyzer": "standard",
      "fielddata": true
    },
    "description": {
      "type": "text",
      "term_vector": "with_positions"
    }
  }
}

Behavior differences

Migration Assistant automatically converts all string fields during metadata migration. No additional configuration is required.

Mixed field types

If your source cluster already contains a mix of string, text, and keyword fields (common in Elasticsearch 5.x with restored Elasticsearch 2.x indexes), only the string fields will be converted. Existing text and keyword fields remain unchanged.

Query and aggregation impacts

After migration, be aware of these key differences:

  • Term queries work differently on text (analyzed) and keyword (exact match) fields.
  • Aggregations: text fields cannot be used for aggregations unless fielddata is enabled; keyword fields are aggregation ready.
  • Sorting: text fields cannot be used for sorting by default; keyword fields support sorting.
  • Case sensitivity: keyword fields are case sensitive; text fields depend on the analyzer.

Application compatibility

Review your application code for:

  • Queries that expect the old string field behavior.
  • Aggregations or sorting on fields now converted to text.
  • Case-sensitive searches on fields now using keyword.

For fields requiring both analyzed and exact-match capabilities, consider using multi-fields with both text and keyword mappings after migration.

Troubleshooting

If you encounter issues with string field conversion:

  1. Verify source version – Ensure your source cluster is running Elasticsearch 1.x through 5.x.

  2. Check migration logs – Review the detailed migration logs for any warnings or errors: 
    tail /shared-logs-output/migration-console-default/*/metadata/*.log
  3. Validate mappings – After migration, verify that the field types have been correctly converted: 
    GET /your-index/_mapping
  4. Review field usage – Ensure that your application queries are compatible with the new field types. Queries that worked with string fields should continue to work with text and keyword fields, but some advanced features may behave differently.
350 characters left

Have a question? .

Want to contribute? or .