I've got a dataframe with a continuous variable x and a grouping factor. I need to add NA as a factor level for some reason. As a result, the data is:

df <- data.frame(x= 1:4, group= factor(c(NA, 1, 1, 2)))
df$group <- addNA(df$group)

How can I now subset the data in the "group" variable for the NA values? I attempted:

df[df$group == "NA", ]
df[df$group == "<NA>", ]
df[is.na(df$group), ]
df[df$group == levels(df$group)[3], ]

My expected output contains all rows where df$group has factor level NA, i.e. data.frame(x= 1:4, group= factor(c(NA, 1, 1, 2)))[1, ].

I do need to add NA as factor level since it is pretty handy in my situation (see here for one case where it is useful).

CodePudding user response：

You can try this approach:

df[is.na(as.character(df$group)),]
  x group
  1  <NA>

moving the variable to character before looking for na values