Return the unique character elements that are mixed with numeric ones in a data.frame

Question

Suppose I have a data.frame that consists of mainly numeric values but also mixed with some character elements.

Is there a way to extract the unique character elements throughout the data.frame?

A toy example along with the desired output is shown below?

   DF <- data.frame(x = c(1:3, "*", "."), y = c("--", 4:6, "="), z = 1:5, w = rep("a", 5))

   desired_output <- c("*", ".", "--", "=")

Answer 1

You could extract all the values which has only punctuations in it from DF using grep :

unique(grep('^[[:punct:]]+$', as.character(unlist(DF)), value = TRUE))
#[1] "*"  "."  "--" "="