Keeping NAs in grepl command

Question

I need to keep NAs in my variables.

I have a variable with names, some have upper and some mixed cases. I want to generate a variable which defines the following name with at least one lower cases = 1 name with only upper cases = 0 NA = (I want to keep it, because I need to make my regressions with non-response)

My code is the following, but I don't know how to implement the NA command.

df <- grepl("[a-z]", x$variable)

Answer 1

You could use ifelse and is.na to check for NA first and only use the result of grepl for values other than NA:

df <- ifelse(is.na(x$variable), NA, grepl("[a-z]", x$variable))

Answer 2

The function below returns

• TRUE if any character is a lower case letter;
• FALSE if no character is a lower case letter;
• NA if the input is NA.

All arguments of grepl can be used, with the same results except for the behavior in the presence of NA's.

na_grepl <- function(pattern, x, ...){
  y <- grepl(pattern, x, ...)
  is.na(y) <- is.na(x)
  y
}

x <- c("abc", "ABc", "ABC", NA)
na_grepl("[a-z]", x)
#[1]  TRUE  TRUE FALSE    NA