0

I'm having trouble finding a solution for a complex variable assignment need. I need to create new variables based on the values in existing variables, but I need to do the same new variable creation many times using a different set of existing variables each time. So for example, I have the following data frame:

dat=data.frame(
kiwi_3=c(1,0,1),
kiwi_5=c(0,0,1),
kiwi_8=c(1,1,0),

apple_3=c(0,0,0),
apple_5=c(1,0,1),
apple_8=c(1,1,0))

I can use the following line of code to create a new variable based on values in existing variables:

dat<-dat %>%
  mutate(fruit_3=case_when(kiwi_3==1 & apple_3==1~1,
                            kiwi_3==1 & apple_3==0~2))

However, what I need to do is to create variables for each number suffix (in this example, fruit_3, fruit_5, fruit_8) based on the variables with the same (non-sequential) number suffix, using the same logic. The actual logic is more complicated than in this example too so I would guess that using case_when will be necessary. I imagine there's a solution using dplyr mutate and across, possibly with indexed vectors with the variable names in them, but I haven't hit upon a solution that works yet.

Thanks for any suggestions!

lemnbalm
  • 5
  • 3

1 Answers1

0

It may be easier to work with column numbers and offsets. Here is a half-baked example using for loops in which the paired values are simply multiplied together to generate the binary results, (assuming that if kiwi = 0 OR apple = 0 the result = 0:

dat=data.frame(
      kiwi_3=c(1,0,1),
      kiwi_5=c(0,0,1),
      kiwi_8=c(1,1,0),
      apple_3=c(0,0,0),
      apple_5=c(1,0,1),
      apple_8=c(1,1,0))

mat <- matrix(NA, nrow = 3, ncol = 3)

for (i in 1:3) {
      for (j in 1:3) {
      mat[i, j] <- dat[i, j] * dat[i, j+3]
      }
}

cbind(dat, mat)
  kiwi_3 kiwi_5 kiwi_8 apple_3 apple_5 apple_8 1 2 3
1      1      0      1       0       1       1 0 0 1
2      0      0      1       0       0       1 0 0 1
3      1      1      0       0       1       0 0 1 0

Column names would also have to be added to the output matrix. This will work on it's own, but others may have ideas on how to refactor the approach using dplyr.

SteveM
  • 2,226
  • 3
  • 12
  • 16
  • Thanks for this response! Unfortunately, the fact that the values can be multiplied were an unintentional result as I simplified my actual case_when function. (I've fixed this above.) Your suggestion of creating a matrix and using a for loop should work though but the code would be very complicated and prone to errors on my part, so I'm wondering if there is a dplyr approach as you suggested. Thanks! – lemnbalm Oct 07 '22 at 15:24