Ingesting a dataset requires one to define a dataset id. This id has to match the "collection" part of the CHO URIS. If they mismatch, the dataset is rejected by Pubby, leading to ingested datasets that cannot be displayed, thus being lost in some kind of intermediate state. This is currently not reported to the user and therefore it would be good to check this at ingestion time.
@kaiec @kba