Is there and R function for recoding variables?

Question

I'm making the leap from SPSS to R but having a few teething issues.

I'm trying to recode a variable but I'm getting error messages.

Here are some examples of the SPSS code I'm trying to translate:

RECODE income (1, 2 = 1) (3, 4 = 2) INTO income2.
EXECUTE.

* Recode to String.
STRING sex_values (A8).
RECODE sex (1 = 'Male') (2 = 'Female') INTO sex_values.
EXECUTE.

Answer 1

Please take a look at a function funnily enough named as recode.

library(dplyr)
char_vec <- sample(c("a", "b", "c"), 10, replace = TRUE)
recode(char_vec, a = "Apple")

Answer 2

Assuming you have a numeric vector containing numbers from the set 1, 2, 3, 4 and you want to replace 1 and 2 with "male" and 3 and 4 with "female" then here are some alternatives.

1) factor This creates such a factor. as.character(income2) can be used if you want a character vector instead.

income <- c(2, 1, 4, 3, 1, 4, 2)
income2 <- factor(income, levels = 1:4, labels = c("male", "male", "female", "female"))
income2
## [1] male   male   female female male   female male  
## Levels: male female

2) subscripting This creates a character vector

income2 <- c("male", "male", "female", "female")[income]
income2
## [1] "male"   "male"   "female" "female" "male"   "female" "male"  

3) car::recode There is a recode function in the car package (and also a slightly different one in the dplyr package and likely others in other packages). This creates a character vector. Add the as.factor = TRUE argument if you want the result to be a factor.

# assumes car package installed
income2 <- car::recode(income, "1:2='male';3:4='female'")
income2
## [1] "male"   "male"   "female" "female" "male"   "female" "male"  

Answer 3

As in many things in R, there are numerous ways to do this.

The first is to use the ifelse function. This is a useful function for reclassifying vectors as it evaluates in a way that conditions are met before replacement so, more complex statments are supported.

( x <- sample(1:2,20,replace=TRUE) )
  ifelse(x == 1, "male", ifelse(x == 2, "female", NA))
  ifelse(x == 1, "male", "female")
  ifelse(x == (1 | 2),"male","female")

You can also directly replace an element in the vector

x[x == 1] <- "male"
x[x == 2] <- "female"
print(x)

There is also using a factor and recoding using levels

( x <- as.factor(sample(1:2,20,replace=TRUE)) )
levels(x)[1] <- "male"
levels(x)[2] <- "female"
print(x)

Answer 4

You may use the if_else() function to replace/recode your data in R for simplicity.

The if_else(condition, true_result, false_result) check if the condition is meet, it returns true_result, if not, returns the false_result instead.

For example x= 3 and if_else(x>2,"yes","no) will return the output:

"yes"

---

So, in your case, you may use the following:

if_else(sex==1,"Male","Female")
if_else(income==(1 | 2),1,2)

In this case you can just put your data's sex column and income column in the above functions to recode them.