How to create new column condition on another columns in R?-CodePudding

I want to create one additional called new column from d,e variables condition on a, b and c columns, but I am struggling to code all these condition in a single function:

#Condition:
    new = e if a=2, b=2 & c=1  
    new = d if a=2, b=1 & c=2  
    new = e if a=2, b=2 & c=2

#data:
df <- data.frame (a  = c(2,2,1,2,1,1),
                  b =  c(2,1,1,2,2,2),
                  c =  c(1,1,2,2,1,2),
                  d =  c(1,3,5,2,7,5),
                  e =  c(4,5,1,7,3,5))

CodePudding user response：

Assuming , means AND:

data.frame(
  a  = c(2, 2, 1, 2, 1, 1),
  b =  c(2, 1, 1, 2, 2, 2),
  c =  c(1, 1, 2, 2, 1, 2),
  d =  c(1, 3, 5, 2, 7, 5),
  e =  c(4, 5, 1, 7, 3, 5)
) %>% 
  mutate(
    new = case_when(
    a == 2 & b == 2 & c == 1 ~ e,
    a == 2 & b == 1 & c == 2 ~ d,
    a == 2 & b == 2 & c == 2 ~ e,
    TRUE ~ NA_real_
    )
  )

Note that the output is:

  a b c d e new
1 2 2 1 1 4   4
2 2 1 1 3 5 NA
3 1 1 2 5 1 NA
4 2 2 2 2 7   7
5 1 2 1 7 3 NA
6 1 2 2 5 5 NA

since the condition does not cover all the cases.

CodePudding user response：

A possible solution:

library(dplyr)

df %>% 
  mutate(new = case_when(a == 2 & b == 2 & c == 1 ~ e,
                         a == 2 & b == 1 & c == 2 ~ d,
                         a == 2 & b == 2 & c == 2 ~ e))

#>   a b c d e new
#> 1 2 2 1 1 4   4
#> 2 2 1 1 3 5  NA
#> 3 1 1 2 5 1  NA
#> 4 2 2 2 2 7   7
#> 5 1 2 1 7 3  NA
#> 6 1 2 2 5 5  NA

CodePudding user response：

data.table option using fifelse:

library(data.table)
setDT(df)[, new := fifelse(a == 2 & b == 2 & c == 1, e,
                   fifelse(a == 2 & b == 1 & c == 2, d,
                   fifelse(a == 2 & b == 2 & c == 2, e, NA_real_)))]

Output:

   a b c d e new
1: 2 2 1 1 4   4
2: 2 1 1 3 5  NA
3: 1 1 2 5 1  NA
4: 2 2 2 2 7   7
5: 1 2 1 7 3  NA
6: 1 2 2 5 5  NA