Description of problem: When I searched for the subfamily Monotropoideae, it brought up the 15 species in the subfamily with iNat observations, but also a few observations of 7 species from Pyroloideae. Both are in the family Ericaceae, but completely different subfamilies. This will happen when i go to exclude a taxon from the search as well. For example, if I were to search “Poaceae” and then exclude Bambusoideae, there might still be a few observations of bamboo included in the search results. I have gone into these odd observations and can’t find anything unusual that would explain why they were showing up in searches for a different taxon. The odd thing is, this only happens sometimes, and other times the search results/species will be completely normal.
Step 1: Search for a taxon above the genus level
Step 2: Go to the species tab, see if all of the species listed belong to the searched taxon. Alternatively, exclude a few daughter taxa (with daughter taxa of their own, such as genera), and see if some of those species are still included.
This is a common indexing bug that can be fixed by inactivating/reactivating the species or adding/removing a DQA vote. Maybe staff will want to take a look, though.
Looks like it was seven out of the eleven most-observed members of Pyroloideae (the four genera of which were recently transferred from Monotropoideae based on this flag): Chimaphila maculata, C. umbellata, Moneses uniflora, Orthilia secunda, Pyrola asarifolia, P. chlorantha, and P. elliptica. (there were only a couple dozen or so observations left behind, though, and all have now been reindexed and are no longer showing up in Monotropoideae search results.)
As for bamboos showing up in Poaceae search results when Bambusoideae is excluded, there are currently 7656 observations of 299 grass species remaining in search results when all grass subfamilies are excluded, including twelve observations of eight bamboo species (Fargesia murielae, F. robusta, Guadua angustifolia, G. paniculata, G. weberbaueri, Neololeba atra, Rhipidocladum racemiflorum, and Yushania microphylla, none of which have a particularly large number of observations or have been moved since the introduction of taxon history). (Some of the grass subfamilies that are organized into tribes also seem to have fair numbers of species-level observations showing up in search results when their relevant tribes are excluded. Much of the subfamilial/tribal/subtribal classification of Poaceae was filled in on iNat in 2020, though some parts of it well predate that—some of the subfamilies, including Bambusoideae for example, were first added several years before then [though they might have been incomplete prior to 2020])
Maybe this helps: I had a similar problem when I tried to retrieve certain taxa via the API. In my case, this happened because a full-text search is being used. It doesn’t only return the scientific names, but also all species that have a matching common name in any language—or even just part of one.
