Elegant and shorter way of recoding

[Slavek]

Hi R Masters,
I have following simple df:

df <- data.frame(
  stringsAsFactors = FALSE,
               URN = c("aaa", "bbb", "ccc"),
               Q9a = c("Exellant", "!", "Honest"),
               Q9b = c("x", NA, "Trustworthy"),
               Q9c = c("Friendly", "Professional", NA)
)

I would like to recode all NAs and short characters into "Blank" and other into "Comment".
I know I can do it using this code:

library(dplyr)
library(stringr)

df <- df %>% 
  mutate(Q9a.Blank = case_when(is.na(Q9a) ~ 1,
                               str_length(Q9a) < 2 ~ 1),
         Q9b.Blank = case_when(is.na(Q9b) ~ 1,
                               str_length(Q9b) < 2 ~ 1),
         Q9c.Blank = case_when(is.na(Q9c) ~ 1,
                               str_length(Q9c) < 2 ~ 1))

df$Q9a.Blank <- recode_factor(df$Q9a.Blank, `1` = "Blank", .missing = "Comment")
df$Q9b.Blank <- recode_factor(df$Q9b.Blank, `1` = "Blank", .missing = "Comment")
df$Q9c.Blank <- recode_factor(df$Q9c.Blank, `1` = "Blank", .missing = "Comment")

But I am sure there is more elegant and much shorter code considering that all variables taken into account have ".Blank" in their names.

Can you help?

[francisbarton]

The new dplyr v1.0.* has support for a new syntax for mutate which might look a bit like this in the second part of your case:

df %>%
  mutate(across(ends_with(".Blank"), ~ recode_factor(`1` = "Blank", .missing = "Comment")))

Answering quickly but I expect you could make something similar for the first half!

[jmcvw]

Maybe something like this

df %>% 
  mutate(across(-URN, ~ifelse(is.na(.) | str_length(.) < 2, 'Blank', 'Comment'))) %>% 
  set_names(c(names(df[1]), paste0(names(df[-1]), '.Blank'))) %>% 
  inner_join(df, .)

[Slavek]

@francisbarton

Answering quickly but I expect you could make something similar for the first half!

Thank you very much but the tricky part in the first part is not simply selecting variables ending with ".Blank" but also selecting relevant variables (a, b or c). I don't know how I could do that

[nirgrahamuk]

library(dplyr)
library(stringr)
df <- data.frame(
  stringsAsFactors = FALSE,
  URN = c("aaa", "bbb", "ccc"),
  Q9a = c("Exellant", "!", "Honest"),
  Q9b = c("x", NA, "Trustworthy"),
  Q9c = c("Friendly", "Professional", NA)
)


caseblank <- function(x){
  case_when(is.na(x) ~ 1,
            str_length(x) < 2 ~ 1)
}


df <- df %>% 
  mutate(across(starts_with("Q9"),
                ~caseblank(.),.names = "{.col}.blank"))

blankf <- function(x){
  recode_factor(x, `1` = "Blank", .missing = "Comment")
}

df %>%  mutate(across(ends_with("blank"), blankf))