Question

Good day,

d <- c(1,1,1,2,2,2,3,3,3)
e <- c(5,6,7,5,6,7,5,6,7)
f <- c(0,0,1,0,1,0,0,0,1)
df <- data.frame(d,e,f)

I have data the looks like the above. What I need to do is for each unique element of d find the first non-zero value in f, and find the corresponding value in e. To be specific, I want another vector g so it looks like this:

d <- c(1,1,1,2,2,2,3,3,3)
e <- c(5,6,7,5,6,7,5,6,7)
f <- c(0,0,1,0,1,0,0,0,1)
g <- c(7,7,7,6,6,6,7,7,7)
df <- data.frame(d,e,f,g)

Suggestions to do this easily? I thought I could use split(), but I am having trouble using which() after the split. I can use ave like this:

foo <- function(x){which(x>0)[1]}
df$t <- ave(df$f,df$d,FUN=foo)

But I am having trouble finding the value of e. Any help is appreciated.

Answer 1

Someone else can provide a base R solution, but here’s a way to do this using plyr:

> ddply(df,.(d),transform,g = head(e[f != 0],1))
  d e f g
1 1 5 0 7
2 1 6 0 7
3 1 7 1 7
4 2 5 0 6
5 2 6 1 6
6 2 7 0 6
7 3 5 0 7
8 3 6 0 7
9 3 7 1 7

Note that I took your note about the “first nonzero element” literally, even though your example data only had a single unique nonzero element in the column (by group).