@Edent @evan the “path_hierarchy” tokenizer makes it much easier to implement general purpose compound query and wildcard machine tag indexing and searching in open/elastic search these days / postgres etc still has the “left most index” problem to support all the possibilities unless something has changed recently – https://github.com/whosonfirst/whosonfirst-opensearch/blob/ee737f8b52501a9fbb69bb48de39f0586c1b36af/schema/2.x/mappings.spelunker.json#L62