Enrich does not allow mapping types with required fields on their definition

[jbaiera]

Enrich now generates a mapping for the enrich index as of 8.9.0 to enable its usage in ES|QL. This process is a simple copy of the mapping type and format, optionally setting index to true if it is supported. Unfortunately, this copying process does not bring along any required configuration values from the mapping's definition.

elasticsearch/x-pack/plugin/enrich/src/main/java/org/elasticsearch/xpack/enrich/EnrichPolicyRunner.java

Lines 387 to 397 in 20e186a

	if (typeAndFormat != null) {
	Map<String, Object> mapping = Maps.newMapWithExpectedSize(3);
	mapping.put("type", typeAndFormat.type);
	if (typeAndFormat.format != null) {
	mapping.put("format", typeAndFormat.format);
	}
	if (isIndexableField(mapperService, enrichField, typeAndFormat.type, mapping)) {
	mapping.put("index", false);
	}
	fieldMappings.put(enrichField, mapping);
	}

This means that a number of mapping types are not able to be included in Enrich:

• aggregate_metric_double
  • Requires metrics and default_metric
• alias
  • Requires path
• scaled_float
  • Requires scaling_factor
• passthrough
  • Requires priority
• semantic_text
  • Requires inference_id
• token_count
  • Requires analyzer

[elasticsearchmachine]

Pinging @elastic/es-analytical-engine (Team:Analytics)

[elasticsearchmachine]

Pinging @elastic/es-data-management (Team:Data Management)

[jbaiera]

Furthermore, the above limitations are just based on the field types that we ship by default in Elasticsearch, referenced via the documentation. If other fields are added via plugins, there's not really a way for us to determine which parameters are supported by the field type when doing this copy operation. It would maybe be beneficial if we had some facility to look up the parameter list and identify which of them are required when transcribing the mappings for Enrich.