The now obsolete anndata package can handle awkward arrays, e.g. adata$obsm$airr <- my_airr where airr is an "awkward.highlevel.Array" object. But in anndataR, this fails with the error:
Error in `private$.validate_aligned_mapping()`:
! Unexpected shape for obsm[['airr']]
ℹ Expected [n], got []
Run `rlang::last_trace()` to see where the error occurred.
where n is the value of nrow(adata). It appears that private$.validate_aligned_mapping does not know how to get the dimensions of airr, whereas the equivalent check in anndata does.
The now obsolete
anndatapackage can handle awkward arrays, e.g.adata$obsm$airr <- my_airrwhereairris an "awkward.highlevel.Array" object. But inanndataR, this fails with the error:where
nis the value ofnrow(adata). It appears thatprivate$.validate_aligned_mappingdoes not know how to get the dimensions ofairr, whereas the equivalent check inanndatadoes.