When doing a tax_glom with a taxrankwhere many taxa are unassigned and NArm = TRUE, then tax_glom() can be bottlenecked by the initial call to prune_taxa() to remove the taxa with bad_empty values at taxrank. I suspect could achieve the same outcome without needing to call prune_taxa() and thereby perhaps gain a significant speed up in such circumstances (especially if taxa_are_rows = TRUE so that the transpose step is skipped).
When doing a
tax_glomwith ataxrankwhere many taxa are unassigned andNArm = TRUE, thentax_glom()can be bottlenecked by the initial call toprune_taxa()to remove the taxa withbad_emptyvalues attaxrank. I suspect could achieve the same outcome without needing to callprune_taxa()and thereby perhaps gain a significant speed up in such circumstances (especially iftaxa_are_rows = TRUEso that the transpose step is skipped).