Why Nominatim stores multiple row of same OSM Object?

[yukha-dw]

As title said. It's strange to see identic OSM Object with different place_id and slightly different class && type values. Is there any advantage of doing this?

Here's the example of duplicated OSM Object. I'm not quite sure if sort_and_cut_results function will always drop the duplicates but it happens in this case:
https://nominatim.openstreetmap.org/search.php?q=%EC%84%9C%EB%9E%98%EC%84%AC&format=jsonv2&debug=1

Query on Postgres

SELECT
 place_id,importance, rank_address, rank_search, osm_type, osm_id, class, type, admin_level, cast(name as json),
  cast(address as json), cast(extratags as json), wikipedia, country_code,
 housenumber, postcode, cast(ST_AsGeoJSON(centroid) as json) as centroid
FROM
 placex
WHERE
 osm_id = '414891746';

place_id	importance	rank_address	rank_search	osm_type	osm_id	class	type	admin_level	name	address	extratags	wikipedia	country_code	housenumber	postcode	centroid
626443	0.25000999999999995	0	20	W	414891746	place	islet	15	{"name": "서래섬", "name:en": "Seorae Island", "name:ja": "ソレ島", "name:ko": "서래섬"}				kr			{"type":"Point","coordinates":[126.990247212,37.50816095]}
626471	0.15000999999999998	0	24	W	414891746	leisure	park	15	{"name": "서래섬", "name:en": "Seorae Island", "name:ja": "ソレ島", "name:ko": "서래섬"}		{"place": "islet"}		kr			{"type":"Point","coordinates":[126.990247212,37.50816095]}

[lonvia]

This is an OSM object that serves two different primary functions. Nominatim duplicates them to be sure they can be found with both functions.