recode() improvements

[GarciaRios]

recode is great! Here are some suggestions:

Character to numeric in factors

recode nicely recodes from character to numeric...So this works:

x <- c("a", "b", "c")
y <- recode(x, a = 1, b = 2, c= 3)
y 

But it won't work with factors even using recode_factor

xf <- factor(c("a", "b", "c"))
yf <- recode(xf, a = 1, b = 2, c= 3)
Error: `a` has type 'double' not 'character'

yf <- recode_factor(xf, a = 1, b = 2, c= 3)
Error: `a` has type 'double' not 'character'

You have to treat it as character and use recode_factor so that it goes back to factor. Seems pretty verbose

recode_factor(as.character(xf), a = 1, b = 2, c= 3)
[1] 1 2 3
Levels: 1 2 3

compatibility with labelled class

  df <- data_frame(s1 = c("M", "M", "F"), s2 = c(1, 1, 2)) %>%
    set_value_labels(s1 = c(Male = "M", Female = "F"), s2 = c(Yes = 1, No = 2))
  
  dplyr::recode(df$s2, `1`=100,`2`=200)

Error in UseMethod("recode") : 
  no applicable method for 'recode' applied to an object of class "labelled"

It would also be great if add_value_labels and remove_value_labels could be incorporated to deal with the unnecessary and new values.

Using car::recode would get us something like this.

library(labelled)
library(car)

  df <- data_frame(s1 = c("M", "M", "F"), s2 = c(1, 1, 2)) %>%
    set_value_labels(s1 = c(Male = "M", Female = "F"), s2 = c(Yes = 1, No = 2))
  
  df$s2 <- car::recode(df$s2, "1=100;2=200")
  df <- df %>% add_value_labels(s2 = c(Yes = 100, No = 200)) %>% 
    remove_value_labels(s2 = c(1,2))
  
  val_labels(df)

[hadley]

I've fixed the first problem: recode() can now recode factors to any data type.

I have not fixed the second problem, as that code should probably go in haven. In the future, can you please file on issue per issue? It makes life easier for me.

[GarciaRios]

Sure, thanks! I will close this issue an post it as a new issue in haven.