OpenSearch: Make Field Queryable Without Creating New Index

gistfile1.txt

### 1 - Update Mapping

```shell
# example of text field

PUT /your_index_name/_mapping
{
  "properties": {
    "your_field_path": {
      "type": "text",
      "fields": { 
        "keyword": { 
          "type": "keyword"
        }
      }
    }
  }
}
```

Result: New documents with this field are immediately queryable. Old documents won't have it, but that's fine.

### 2 - Update existing documents

```shell
POST /your_index_name/_update_by_query?wait_for_completion=false&conflicts=proceed
{
  "query": {
    "match_all": {}
  }
}
```

Optional - Use a script to populate default values:

```shell
POST /your_index_name/_update_by_query
{
  "query": {
    "match_all": {}
  },
  "script": {
    "source": "if (ctx._source.my_field == null) { ctx._source.my_field = 'default_value' }"
  }
}
```

### 3 - Refresh

```shell
POST /your_index_name/_refresh
```

# create pipeline
PUT _ingest/pipeline/core_products_default_pipeline
{
  "processors": [
    {
      "script": {
        "lang": "painless",
        "source": """
          def parts = new ArrayList();

          if (ctx.sku != null && ctx.sku.toString().trim().length() > 0) {
            parts.add(ctx.sku.toString().trim().toLowerCase());
          }

          if (ctx.name != null && ctx.name.toString().trim().length() > 0) {
            parts.add(ctx.name.toString().trim().toLowerCase());
          }

          if (ctx.barcode != null && ctx.barcode.toString().trim().length() > 0) {
            parts.add(ctx.barcode.toString().trim().toLowerCase());
          }

          if (ctx.upc != null && ctx.upc.toString().trim().length() > 0) {
            parts.add(ctx.upc.toString().trim().toLowerCase());
          }

          if (ctx.gtin != null && ctx.gtin.toString().trim().length() > 0) {
            parts.add(ctx.gtin.toString().trim().toLowerCase());
          }

          if (!parts.isEmpty()) {
            ctx.product_search_combined = String.join(' ', parts);
          } else {
            ctx.product_search_combined = '';
          }
        """
      }
    }
  ]
}

# list all pipeline
GET _ingest/pipeline

# pipeline details
GET _ingest/pipeline/core_products_default_pipeline

# update mapping -> need refresh mapping & fields
PUT core_products/_mapping
{
  "properties": {
    "product_search_combined": {
      "type": "keyword",
      "ignore_above": 32766
    }
  }
}

# set default pipeline for index, new doc will have new field
PUT core_products/_settings
{
  "index.default_pipeline": "core_products_default_pipeline"
}

# update existing docs
# NOTE: Make sure free disk > used disk
POST core_products/_update_by_query?pipeline=core_products_default_pipeline&conflicts=proceed&wait_for_completion=false
{
  "query": {
    "match_all": {}
  }
}

# view task progress, task id get from prev action
GET _tasks/1zsw2RngShWru4A4sMJ3UA:78005101

# refresh index & fields by API or manual refresh
POST core_products/_refresh