Recode variable in R based on first characters of vector

Question

I have a data frame which looks like this:

codes <- c('TFAA1', 'TFAA2', 'TFAA3', 'TFAA4', 'TFAB1', 'TFAB2', 'TFAB3', 'TFAB4')
scores <- c(4,3,2,2,4,5,1,2)
example <- data.frame(codes, scores)

I want to create a new column called code_group whereby everything that starts with TFAA gets called "Group1" and everything that starts with TFAB gets called "Group2".

Have been playing with the recode function from the car package, and the grepl function but I'm failing miserably. Here's my attempt so far....

recode <- (codes, "%in% TFAA='Group1'; %in% TFAB='Group2'")

RLave · Accepted Answer · 2018-07-19T14:41:46.113

3

With dplyr and stringr you can get it done:

library(dplyr)
library(stringr)
example %>% 
  mutate(code_group = case_when(str_detect(codes, "^TFAA") ~ "Group1",
                              str_detect(codes, "^TFAB") ~ "Group2"))

case_when lets you use multiple if-then cases. str_detect lets you, well, detect the pattern you seek in a string.

edited Jul 19 '18 at 14:41

answered Jul 19 '18 at 14:19

RLave

8,144
3
21
37

`startsWith` can substitute `str_detect` of course. – RLave Jul 19 '18 at 14:41

score 1 · Answer 2 · answered Jul 19 '18 at 14:35

1

example$code_group <- ifelse(startsWith(codes, 'TFAA'), 'Group 1', 
                      ifelse(startsWith(codes, 'TFAB'), 'Group 2',
                             NA))

answered Jul 19 '18 at 14:35

IceCreamToucan

28,083
2
22
38

score 1 · Answer 3 · answered Jul 19 '18 at 15:45

We could extract the first four characters with substr, convert it to factor and specify the labels as the one we wanted

example$code_group <-  with(example,  as.character(factor(substr(codes, 1, 4), 
              levels = c('TFAA', 'TFAB'), labels = c('Group1', 'Group2'))))

moodymudskipper · Answer 4 · 2018-07-20T00:38:05.247

We can use split<- :

example$group <- NA
split(example$group,substr(example$codes,1,4)) <- paste0("Group",1:2)
example
#   codes scores  group
# 1 TFAA1      4 Group1
# 2 TFAA2      3 Group1
# 3 TFAA3      2 Group1
# 4 TFAA4      2 Group1
# 5 TFAB1      4 Group2
# 6 TFAB2      5 Group2
# 7 TFAB3      1 Group2
# 8 TFAB4      2 Group2

Or we can use factors for the same output (3 variants):

example$group <- paste0("Group",factor(substr(example$codes,1,4),,1:2))
example$group <- paste0("Group",as.numeric(factor(substr(example$codes,1,4))))
example$group <- factor(substr(example$codes,1,4),,paste0("Group",1:2))

In the last case you get a factor column, in all other cases you get a character column.

Recode variable in R based on first characters of vector

4 Answers4