dyplr / grouped_df and implicit NA warning

tidyverse
1

We have code that recently started to give the following warnings whenever we subset a grouped dataframe that results in zero rows. I understand what 'Factor x contains implicit NA' refers to, and understand that I could use fct_explicit_na to fix the warning; however, this doesnt seem correct to report it in the situation of zero rows (as opposed to actually having NAs). I only see this with grouped dataframes. See this example:

library(dplyr)

#Create DF:
test <- data.frame(x = factor(c('A', 'B')))

#Group by make grouped_df:
test <- test %>% group_by(x)

#Subset, resulting in zero rows:
ss <- test[test$x == 'C',]

# Warning message:
# Factor `x` contains implicit NA, consider using `forcats::fct_explicit_na`

I am using dplyr_0.8.3. Thanks for any comments.

2

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.