Using subsets and whole dataframes simultaneously in a loop - r

Question

I'm trying to write a function that loops over rows of a dataframe and uses information about other rows to determine the output for each loop.
Consider the following dataframe, which is meant to represent people who have a longitude coordinate, a latitude coordinate, and a value to represent if they are or are not sick:
game.mat<-as.data.frame(matrix(0, nrow = 100, ncol = 3))
colnames(game.mat)<-c("PosX","PosY","Sick")
game.mat[,"PosX"]<-sample(x = c(1:100), 100, replace = TRUE)
game.mat[,"PosY"]<-sample(x = c(1:100), 100, replace = TRUE)
game.mat[,"Sick"]<-sample((c(rep(0,8),rep(1,2))),100,replace=TRUE)
Some minority of people will be sick at baseline. My function is meant to infect people who have neighboring x-y coordinates with a sick person (so anyone who is adjacent to someone who is sick). I considered embedding a function like this in an ifelse statement:
search_sick<-function(d,corx,cory){
d2<-d[d$PosX<corx+2&d$PosX>corx-2&d$PosY<cory+2&d$PosY>cory-2,]
if(sum(d2$Sick>0)){
d$Sick<-1
} else{
d$Sick<-0
}
}
But it makes everyone sick, perhaps because it gives everyone a value of 1 if anyone is next to a sick person. I also considered using an apply function. But from what I understand about apply, it will only operate within the a single row at a time so it will be impossible to retrieve information about whether other rows have neighboring coordinate values.
I hope this makes sense. Happy to provide any additional information.

Answer 1

Here's an example using apply
set.seed(1)
game.mat<-as.data.frame(matrix(0, nrow = 100, ncol = 3))
colnames(game.mat)<-c("PosX","PosY","Sick")
game.mat[,"PosX"]<-sample(x = c(1:100), 100, replace = TRUE)
game.mat[,"PosY"]<-sample(x = c(1:100), 100, replace = TRUE)
game.mat[,"Sick"]<-sample((c(rep(0,8),rep(1,2))),100,replace=TRUE)
#plot the sick individuals in red
plot(PosY~PosX, data=game.mat, col=as.factor(Sick), pch=16)
We'll modify your function to have a flexible search radius "r", and to return the indices of the newly infected individuals
search_sick<-function(d, corx, cory, r){
indx<-which(d$PosX<corx+r & d$PosX>corx-r & d$PosY<cory+r & d$PosY>cory-r)
}
contagious<-game.mat[game.mat$Sick==1,]
infected<-apply(contagious, 1, function(x) {
search_sick(game.mat, x[1], x[2], r=10)
})
game.mat$T1<-game.mat$Sick
game.mat$T1[unique(unlist(infected))]<-1
#circle points which have become sick
points(PosY~PosX, data=game.mat[game.mat$Sick==0 & game.mat$T1==1,], col="red", cex=2)