combine_cols_blanks.R

library(dplyr)
library(tidyr)

words_some_missing <- data_frame(one_word = c(NA, "M", NA), 
                                 second_word = c("F", NA, NA))

words_some_missing %>% 
  unite(sex, one_word, second_word)

## that sticks the NAs together, which I think is something you dont want? In 
## that case, just replace the NAs with and empty string before combining them.
## Then, combine them with another empty string as replacement:

words_some_missing %>% 
  replace_na(list(one_word = "", second_word = "")) %>% 
  unite(sex, one_word, second_word, sep = "")

# I'm assuming here that this is what you want -- you have the sex spread across
# several columns and you want to bring it into a single column?