COL XR seems to include a lot of duplicate records for the same taxa, especially in the ranks of family and genus, and those in between. For example, there are three accepted genera called Chrysis within the order Hymenoptera, with different levels of taxonomic info, different constituent species, and importantly (very) different parent taxa. See https://github.com/CatalogueOfLife/data/issues/1637 for more info there. Because the main genus record is erroneously placed in Vespidae, 62,584 occurrence records (out of ~190,000) no longer show up when filtering by the family Chrysididae.
Does the algorithm used to combine taxonomies from different sources work differently in COL XR than in the GBIF Backbone Taxonomy? How could I go about curating these issues (without, in the case of this example, maintaining a database of all Chrysididae)?
I also wonder, if this is an issue with the merging algorithm, does this affect decisions to add other sources to COL XR? I was planning to submit checklists from my database for inclusion in COL XR but if the current setup has issues with merging taxonomies from heterogeneous sources that might be counterproductive.