How to REMOVE lower case tokens with R?

Question

I'm using R/Quanteda and I'm trying to make a wordcloud from ONLY upper case words. The txt is a from a bibliographic reference in ABNT format, doing so I would keep only the authors surnames. Any hint? Tanks!

Answer 1

Base R

string <- "lowercase UPPERCASE more lower case UPPER  1143 + 40 = !!!"

gsub(" {2,}", " ", # replace 2 or more consecutive spaces with one space
  gsub("[^A-Z ]", "", string) # remove anything that's not a space or an uppercase letter
)
#> [1] " UPPERCASE UPPER "

^{Created on 2021-07-02 by the reprex package (v2.0.0)}

Stringr Package

require(stringr)
#> Loading required package: stringr

string <- "lowercase UPPERCASE more lower case UPPER  1143 + 40 = !!!"

str_squish( # remove excess whitespace
  str_remove_all(string, "[^[:UPPER:] ]") #remove everything except uppecase and spaces
)
#> [1] "UPPERCASE UPPER"

^{Created on 2021-07-02 by the reprex package (v2.0.0)}