[BenjaminDEMAILLE]

Fixes #256.

> RenameCells(object, new.names = rep("dup", ncol(object)))
Error in `.rowNamesDF<-`(x, value = value): duplicate 'row.names' are not allowed

The mapping is applied to the meta data, the identities and every assay in turn, so the failure comes from whichever frame is renamed first. The message names neither the cells nor the renaming, and the traceback lands in .rowNamesDF<-, which is why that issue reads as a mystery.

Checking the mapping before applying it:

Renaming would give the same name to more than one cell. 1 name is repeated: 'dup'

Cell names have to be unique, so this was never going to succeed; it just says so now, and names what to fix.

The other half of that issue

The reporter also hits it through RunAzimuth(), with a different message (missing values in 'row.names' are not allowed) and a different cause: an SCT model naming cells the assay no longer has. That one is satijalab/seurat#10481, since the code is in Seurat. It also explains why passing make.unique(Cells(object)) did not help them: uniqueness was not the problem in that path.

Tests

tests/testthat/test_rename_duplicates.R: every cell renamed to one name, a single collision named exactly, and unique names plus add.cell.id still renaming.

Three assertions fail without the change. SeuratObject suite: 182 passing, 0 failures. Seurat's own suite against this build: 1184 passing, the only 2 failures being ones that predate it and are unrelated (Read10X_Image, satijalab/seurat#10454).

🤖 Generated with Claude Code

[BenjaminDEMAILLE]

Superseded by #319, which groups the message-only PRs into one, since they share a rationale and none of them changes behaviour. The commit is carried over unchanged; closing here to keep the queue down rather than because anything was wrong with it.