使用不同的列应用功能-R

Question

I have an array in r that has thousands of rows and three different columns. For each row in the array, I want to change the first column based on the first column and third column. Specifically, if the number in the third column is larger than the number in the first column, it should return a 1, if the number in the third column is smaller than the number in the first column, it should return a 0, if they are the same number it should randomly choose 1 or 0.

    temp[,1]<- sapply(
                        temp[,1],
                        function(score){
                           if(temp[,3]>score){1}
                           else if(temp[,3]<score){0}else
                           else if(temp[,3]==score){sample(0:1,1)}})

This is the code I have so far but it isn't quite working, I think when I call "temp[,3]" it returns a vector with all the rows in the third column, I just want the one entry in that row corresponding to the row I get with "score".

Answer 1

也可以使用算术代替条件来完成：

rbinom(nrow(temp), 1, (temp[,1]<temp[,3])+(temp[,1]==temp[,3])*0.5)

Answer 2

Use ifelse :

ifelse(temp[,1] < temp[,3], 1, # test condition 1, give 1 if TRUE
       ifelse(temp[,1] > temp[,3], 0, # condition 1 is FALSE, test condition 2
              rbinom(dim(temp)[1],1,.5))) # condition 2 is FALSE, return random

EDIT : ifelse works by testing a condition, then doing something if that condition is TRUE or something else if it is FALSE . You have three conditions, so instead of providing a value for when the first condition ( temp[,1] < temp[,3] ) is FALSE , this provides a second ifelse statement, which tests your second condition (temp[,1] > temp[,3]) .

All remaining cases are then generated randomly using rbinom . This has to be a vector rather than draw because otherwise all of your rows that do not meet conditions 1 and 2 would be given the same random draw. If you try:

ifelse(temp[,1]<temp[,3],1,ifelse(temp[,1]>temp[,3],0,rbinom(1,1,.5)))

You'll see that all rows where temp[,1]==temp[,3] have the same value. It would be the same as just doing something like:

ifelse(temp[,1]<temp[,3],1,ifelse(temp[,1]>temp[,3],0,2))