On PC-AiR (if using group variable) and run null model steps, I've noticed an error is thrown if the sample.id's in the annotated dataframe are integers. I believe something in the background is converting these ids to character, and then when a subsequent join is attempted it fails because the object type of the join column (sample.id) is inconsistent. Explicitly setting this column to character when constructing the annotated DF appears to fix the problem.
On PC-AiR (if using group variable) and run null model steps, I've noticed an error is thrown if the sample.id's in the annotated dataframe are integers. I believe something in the background is converting these ids to character, and then when a subsequent join is attempted it fails because the object type of the join column (sample.id) is inconsistent. Explicitly setting this column to character when constructing the annotated DF appears to fix the problem.