How do I calculate Euclidean distances across NA values in r - r

Question

I have a date frame like this
individual <- c("1",NA,NA,NA,NA,NA,NA,NA,"1","1")
x <- c(665,NA,NA,NA,NA,NA,NA,NA,663,665)
y <- c(-474.5,NA,NA,NA,NA,NA,NA,NA,-474.5,-472.5)
frame <- rep(1:10)
df <- data.frame(individual,x,y,frame)
I have an ID column labeled 'individual', xy coordinates, and a frame number.
I need to calculate the euclidean distances for the x,y coordinates between rows but over the NA values.
So, in the example I gave - I would need to calculate the distances between rows 1 and 9, as well as 10 and 9. In the real data there would be substantially more rows of course.
Eventually what I need to do is interpolate the data, so that if the euclidean distance is <5, fill in the data rows that are missing with the ID of the individual. If the euclidean distance is >5, then ignore and interpolate nothing.
Here is the example result data frame that's needed:
individual <- c("1","1","1","1","1","1","1","1","1","1")
x <- c(665,NA,NA,NA,NA,NA,NA,NA,663,665)
y <- c(-474.5,NA,NA,NA,NA,NA,NA,NA,-474.5,-472.5)
frame <- rep(1:10)
dist_measure <- c(NA,NA,NA,NA,NA,NA,NA,NA,2,2.828427)
df <- data.frame(individual,x,y,frame,dist_measure)
Any advice on an approach to this problem is greatly appreciated. My first thought was to have a function that calculates Euclidean distance and put it in a for loop. But I'm a bit stuck on how to work this over the NA values. I thought somehow using the lag function in the tidyverse would help, but not sure again how to integrate that into the loop/function.
Thank you in advance.

Answer 1

This should work. I've added another individual into the hypothetical data to show how it works.
individual <- c("1",NA,NA,NA,NA,NA,NA,NA,"1","1",
"2",NA,NA,NA,NA,NA,NA,NA,"2","2")
x <- c(665,NA,NA,NA,NA,NA,NA,NA,663,665,
.665,NA,NA,NA,NA,NA,NA,NA,.663,.665)
y <- c(-474.5,NA,NA,NA,NA,NA,NA,NA,-474.5,-472.5,
-.4745,NA,NA,NA,NA,NA,NA,NA,-.4745,-.4725)
frame <- rep(1:10, 2)
df <- data.frame(individual,x,y,frame)
for(i in 1:2){
tmp <- df[min(which(df$individual == as.character(i))):
max(which(df$individual == as.character(i))), ]
ends <- range(which(is.na(tmp$individual))) + c(-1,1)
if(nrow(tmp) > 1 & ends[1] > 0 & ends[2] <= nrow(tmp)){
d <- c(dist(tmp[ends, c("x", "y")]))
if(d < 5){
df$individual[min(which(df$individual == as.character(i))):
max(which(df$individual == as.character(i)))] <- tmp$individual[ends[1]]
}
}
}
df
# individual x y frame
# 1 1 665.000 -474.5000 1
# 2 1 NA NA 2
# 3 1 NA NA 3
# 4 1 NA NA 4
# 5 1 NA NA 5
# 6 1 NA NA 6
# 7 1 NA NA 7
# 8 1 NA NA 8
# 9 1 663.000 -474.5000 9
# 10 1 665.000 -472.5000 10
# 11 2 0.665 -0.4745 1
# 12 2 NA NA 2
# 13 2 NA NA 3
# 14 2 NA NA 4
# 15 2 NA NA 5
# 16 2 NA NA 6
# 17 2 NA NA 7
# 18 2 NA NA 8
# 19 2 0.663 -0.4745 9
# 20 2 0.665 -0.4725 10