Avoiding nested ifelse without using "dplyr" library

Question

Actually, I want to assign to each credit_category a specific Risk-weight.

I've been using nested ifelse as shown below, but I am trying to find something else in order to avoid the nesting, without using "dplyr" library. Any ideas ?

tab_nonsec$RW = ifelse(tab_nonsec$credit_category=="AAA", 0.005,
                       ifelse(tab_nonsec$credit_category=="AA", 0.02,
                       ifelse(tab_nonsec$credit_category=="A", 0.03,
                       ifelse(tab_nonsec$credit_category=="BBB", 0.06,
                       ifelse(tab_nonsec$credit_category=="BB", 0.15,
                       ifelse(tab_nonsec$credit_category=="B", 0.3,
                       ifelse(tab_nonsec$credit_category=="CCC", 0.5,
                       ifelse(tab_nonsec$credit_category=="Unrated", 0.75,
                       ifelse(tab_nonsec$credit_category=="Defaulted",1,0
                       )))))))))

Or create a [lookup table](https://stackoverflow.com/questions/8433523/creating-a-new-variable-from-a-lookup-table) first and then merge. — markus, Jun 07 '19 at 09:18
@Sotos OP is mapping character to numeric not the other way around... still **definitely** a duplicated question. — s_baldur, Jun 07 '19 at 09:24
If you have many categories, `merge` would be more suitable. — zx8754, Jun 07 '19 at 09:29

score 1 · Accepted Answer · edited Jun 07 '19 at 09:25

1

car::recode(test, '"AAA"=0.005;
                  "AA"=0.02;
                  "A"=0.03;
                  "BBB"=0.06;
                  "BB"=0.15;
                  "B"=0.3;
                  "CCC"=0.5;
                  "Unrated"=0.75;
                  "Defaulted"=1')

The recode function is pretty easy to use.

edited Jun 07 '19 at 09:25

NelsonGon

13,015
7
27
57

answered Jun 07 '19 at 09:23

DSGym

2,807
1
6
18

If I had many categories to recode I would probably try to get the cateogories in a csv and use the vectorized `mutate` function to recode my data :-) – DSGym Jun 07 '19 at 09:27

score 0 · Answer 2 · answered Jun 07 '19 at 09:22

You can used a named vector to make a lookup table:

lut <- c(
  "AAA" = 0.005, "AA" = 0.02, "A" = 0.03, 
  "BBB" = 0.06, "BB" = 0.15, "B" = 0.3, 
  "CCC" = 0.5, 
  "Unrated" = 0.75, 
  "Defaulted" = 1
)

tab_nonsec$RW <- lut[tab_nonsec$credit_category]
tab_nonsec$RW[is.na(tab_nonsec$RW)] <-0

#   credit_category    RW
# 1              AA 0.020
# 2             CCC 0.500
# 3             AAA 0.005
# 4               B 0.300
# 5             AAA 0.005
# 6       Defaulted 1.000
# 7             BBB 0.060
# 8             BBB 0.060
# 9          errror 0.000

# Example data:

tab_nonsec <- data.frame(credit_category = c("AA", "CCC", "AAA", "B", "AAA", "Defaulted", "BBB", "BBB", "errror"), stringsAsFactors = FALSE)

Avoiding nested ifelse without using "dplyr" library

2 Answers2