condition has length > 1 and only the first element r

Question

I have

df = data.frame(Col1 = c( NA, 1," ", 2345.768,"hi","", NA, 3.4, "44.99"))

and want to format specific values, and created a udf

format_it = function(y, n_decimals, dash_type, suffix = ""){
  if(is.na(y)) return(dash_type)
  if(nchar(gsub(" ", "", y))==0) return(y)
  has_letter = grep("[A-z]+", y)
  if(is_empty(has_letter)== TRUE) {
    return(paste0(format(round(as.numeric(y), n_decimals), nsmall=n_decimals, big.mark = ","),suffix))
  }
  if(has_letter == 1){ 
    return(y)
  } else{
    x = as.numeric(y)
    ifelse(is.na(x), 
         dash_type,
         paste0(format(round(as.numeric(x), n_decimals), nsmall=n_decimals, big.mark = ","),suffix))}
  
}

I tested each value individually, ie format_it(df$Col1[1],1,"-") , and each one worked ok

but, when I set up a set_formatter in flextable,

df %>%
  flextable() %>%
  set_formatter(Col1 = function(x) format_it(x,1,"-"))

I hoped the results would be correct, but received the wrong results,

with the message: the condition has length > 1 and only the first element will be used

I tried updating to include Vectorize , but received the same error

Any suggestions?

I would like to see

Answer 1

I'm a little confused on your function, but a fresh code approach to recreating your table (based on your function) in a reproducible way is below, which produces your desired output. It first replaces any NA values in the original data with "-", then checks for all non-numeric values (ie, "hi") using grepl and keeps those the same, then standardizes the significant digits in the numeric values with sprintf . This approach was within the dplyr "world" using mutate() and case_when() and did not use a user-defined function.

df %>% 
  mutate(Col1 = case_when(
    is.na(Col1) ~ "-",
    !grepl("[^A-Za-z]", Col1) ~ Col1,
    grepl(".", Col1) ~ sprintf("%.1f", as.numeric(Col1)),
  )) %>%
  flextable::flextable()