DADA2_SPLITREGIONS fails with only one sample

[julianSchwab]

Description of the bug

I know this is an unusual case, but I was running the pipeline for a benchmark only using one sample. However, DADA2_SPLITREGIONS fails with the following error:
Error in rowSums(df[, 2:(ncol(df) - 1)] != 0) : 'x' must be an array of at least two dimensions

The reason is that for only one sample the command df <- df[as.logical(rowSums(df[,2:(ncol(df)-1)] != 0)), ] fails due to automatic matrix to vector conversion if only one column is selected. Thus, rowSums produces this error.

For the sake of robustness, I would suggest to escape this rare case by e.g. enforcing matrix-structure via df[as.logical(as.matrix(rowSums(df[,2:(ncol(df)-1)] != 0))), ]

Command used and terminal output

Relevant files

No response

System information

No response

[d4straub]

One sample is indeed unusual but we should make it work regardless. Thanks for suggesting a fix! Could you open a PR with the suggested changes?