[BenjaminDEMAILLE]

Two ways DotPlot() fails on input it should handle. Both found by sweeping the plotting functions with degenerate but ordinary arguments; VlnPlot() and FeaturePlot() handle the same inputs.

A feature named twice

> DotPlot(object, features = c(markers, markers[1]))
Error in `levels<-`: factor level [4] is duplicated

From

data.plot$features.plot <- factor(x = data.plot$features.plot, levels = features)

The features are there only to fix the order of the axis, so they are taken unique. This happens as soon as marker lists are combined, and with a named list (the grouped dot plot) it is unavoidable when a gene marks two groups.

Cells with no identity

> DotPlot(object, features = markers)
Error in `rbind`: duplicate 'row.names' are not allowed

CellsByIdentities() returns a group for NA, padded with NA cell names, and those reach FetchData(). The per-group frames then fail to bind. Such cells are now dropped, with

Warning: 3 cells have no identity and are not shown

Both the cell names and the identity column are filtered, so this holds whether or not the padding itself is fixed. With satijalab/seurat-object#302 (which fixes the padding) but without this change, the same call still fails, differently:

Error: replacement has 1 row, data has 0

so the two are complementary rather than duplicates.

Tests

tests/testthat/test_dotplot_features_idents.R: a repeated feature plotted once, a feature repeated across named groups, cells with no identity warned about and excluded, and the ordinary case unchanged.

Three assertions fail without the change. Full suite NOT_CRAN=true: 1188 passing, the only 2 failures being the pre-existing Read10X_Image ones fixed in #10454.

🤖 Generated with Claude Code

[BenjaminDEMAILLE]

Superseded by #10492, which groups five small independent fixes into one review. The commit is carried over unchanged; closing here to keep the queue down rather than because anything was wrong with it.