CODEWITHSUNDEEP
rdata.table
bvowe
bvowe

Reputation: 3384

R DataTable Two FifElse Statement

    data = data.frame("var1" = c("one", "two", "three", "four", NA))



 I seek a data.table fifelse solution so get, 

data = data.frame("var1" = c("one", "two", "three", "four", NA),
"var2" = c(10,20,30,40,NA)

The rule is: if var1 = "one" var2 = "10"

if var1 = "two" var2 = "20"

if var1 = "three" var2 = "30"

if var1 = "four" var2 = "40"

And then I hope to label '10', '20', '30', '40' as : dog, cat, fox bird

So how to use multiple fifelse statement and label variable in data.table

Upvotes: 1

Views: 885

Answers (1)

akrun
akrun

Reputation: 887531

If we need to create 'var2', we can pass a named vector to match and create the new 'var2'

library(data.table)
setDT(data)[, var2 := setNames(c(10, 20, 30, 40),
        c('one', 'two', 'three', 'four'))[as.character(var1)]]
data[, label := setNames(c('dog', 'cat', 'fox', 'bird'), c(10, 20, 30, 40))[as.character(var2)]][]  


data
#    var1 var2 label
#1:   one   10   dog
#2:   two   20   cat
#3: three   30   fox
#4:  four   40  bird
#5:  <NA>   NA  <NA>

It may be more efficient to do this with a named vector, but fifelse can be used as well

setDT(data)[, var2 := fifelse(var1 == 'one', 10,
                     fifelse(var1 == 'two', 20,
                     fifelse(var1 == 'three', 30,
                     fifelse(var1 == 'four', 40, NA_real_))))]

and similarly, for 'label'

Or in the devel version, there is fcase

setDT(data)[, var2 := fcase(var1 == 'one', 10,
                             var1 == 'two', 20,
                             var1 == 'three', 30,
                             var1 == 'four', 40)]

Upvotes: 1

Related Questions