从 R 数据框中的多列创建新变量的有效方法

Question

I'm trying to create a new variable called DRG from a set of 480 variables based on some condition, the new variable is a binary if the condition is true. if any of the columns in the data frame has the values 060 or 191 then DRG =1 else DRG =0;

 for (i in 1:nrow(DATA_opioid)){

   for (j in 42:480)

     { if (!is.na(DATA_opioid[i,j])  {

     if ( (DATA_opioid[i,j]) == '060' | (DATA_opioid[i,j]) == '191'| (DATA_opioid[i,j+1]))==           
    '060' |(!is.na(DATA_opioid[i,j+1]))=='191')

        { 
          DATA_opioid$DRG =1
        }
      else DATA_opioid$DRG =0

       }
   }

I have been unable to get arrive at a working code although I did succeeded when I tried it for one of the columns. but there are 480 variables of all starting with prefix 'RX'. Any useful suggestion to solve this is most welcome.

for (i in 1:nrow(DATA_opioid)){
    if (DATA_opioid$RX1CAT1[i]  == "060" | DATA_opioid$RX1CAT1[i] == "191"){

    DATA_opioid$DRG[i] =1 

}
else DATA_opioid$DRG[i] =0
}

Answer 1

You don't need to use loops for such operations. There are many ways to do this. Here are few one.

Using rowSums

df$DRG <- +(rowSums(df == '191' | df == '060') > 0)

#    a   b DRG
#1   1   2   0
#2   2   3   0
#3   3   4   0
#4   4 060   1
#5   5   3   0
#6 191   4   1

Using apply

df$DRG <- +(apply(df == '191' | df == '060', 1, any))

We can also use rowSums in dplyr chain

library(dplyr)
df %>% mutate(DRG = +(rowSums(. == '191' | . == '060') > 0))

---

If you want to test this only on some columns subset the dataframe for those columns in the above solution. For example to test for columns 3 to 5 you can do

df$DRG <- +(apply(df[3:5] == '191' | df == '060', 1, any))

data

Tested on this data :

df <- data.frame(a = c(1:5, 191), b = c(2:4, '060', 3:4))

Answer 2

assuming your dataframe is called df:

DRG<-apply(df,1,function(x){
  max(x == "060" | x == 191)
})