Easy way of executing changes on an existing Elastic Search Mapping through code. This is very similar to Database Migration mechanisms such as DB Migration package.

yarn add elasticsearch-mapping-migration

Then run

yarn esm-migration help

Set the following environment variables

ESMIGRATION_HOST=your-elasticsearch-host
ESMIGRATION_PORT=your-elasticsearch-port
ESMIGRATION_PROTOCOL=your-elasticsearch-protocol-http-or-https
ESMIGRATION_MIGRATION_LIST_FILE=path/to/your/migration-list.json
ESMIGRATION_MIGRATION_INDEX=esm-migration

All paths must be relative to your project root directory. So no absolute path here.

The variable ESMIGRATION_MIGRATION_LIST_FILE set where your list of migrations will be.

the variable ESMIGRATION_MIGRATION_INDEX set which elastic search index will be created to keep track of all migrations.

yarn esm-migration init

This will create an index to store all performed migrations identified by timestamp

If the init is run again then you have

Migration list example

[
  {
    "timestamp": "20200612170400",
    "file": "src/migrations/migration-1.json"
  },
  {
    "timestamp": "20200612170401",
    "file": "src/migrations/migration-2.json"
  }
]

The migration list file contains a list of migrations to be executed. Each item has the property timestamp which is an unique identifier.

Once it is performed, then that timestamp is stored. That is how a migration is only executed once.

Migration file example

{
  "index_target": "<YOUR-INDEX-NAME>",
  "type": "<YOUR-FIELD-NAME>",
  "migration" : {
    "properties": {
      "<YOUR-FIELD-NAME>": {
        "type": "text",
        "fielddata": true,
        "fields": {
          "raw": {
            "type": "keyword"
          },
          "normalized": {
            "type": "keyword",
            "normalizer": "name_keyword_normalizer"
          }
        }
      }
    }
  }
}

As Elastic Search does not allow multiple field mapping update, every migration file will only have 1 field update

index_target is the index your documents are and where the mapping update will be performed on type is the field name migration is the actual mapping update. See Elastic Search Update Mapping Reference

yarn esm-migration migrate

Then there should be an output like below

When the command esm-migration migrate is executed, the migration list file is loaded. Then all migrations within will be performed.