Calculating the difference between two values in the same column by ID based one the value of another column in R

Question

I have a dataset that looks like this:

threadid    unixtime    sent    ID  
123         1000        0       1   
123         1050        1       1   
123         1070        0       1   
123         2000        1       1   
123         2500        1       1   
123         3000        0       1   
123         1000        0       2   
123         1500        0       2   
123         2500        1       2  

But I want it to look like this:

threadid    unixtime    sent    ID  change
123         1000        0       1   
123         1050        1       1   
123         1070        0       1   
123         2000        1       1   
123         2500        1       1   1430
123         3000        0       1   
123         1000        0       2   
123         1500        0       2   
123         2500        1       2   1000

So by ID, I want to look for the last occurrence of a "1" and then calculate the time difference between the unix time that corresponds with the 1 and the previous observation (or last observation with a 0 in the "sent" column) that has 0 in the "sent" column. I think this may involve a "for" loop but I've tried a lot of things and just can't quite get it. Any help is greatly appreciated!

Answer 1

This is probably not the most efficient way to do this but you could try:

  library(dplyr)
  getDiff<-function(x){
        x$change<-''

        if(sum(unique(x$sent)==c(0,1))==2){
        #get the max of the indexes where sent==1
        lastSent<-max(which(x$sent==1))

        #get the max of the indexes where sent==0 and that are smaller than lastSent      
        lastBeforeSent<-max(which(x$sent==0)[which(x$sent==0)<lastSent])

        x$change[lastSent]<-x$unixtime[lastSent]-x$unixtime[lastBeforeSent]
        }
        return(x)
}

Run on your data it gives:

  threadid unixtime sent ID change
1      123     1000    0  1       
2      123     1050    1  1       
3      123     1070    0  1       
4      123     2000    1  1       
5      123     2500    1  1   1430
6      123     3000    0  1       
7      123     1000    0  2       
8      123     1500    0  2       
9      123     2500    1  2   1000