How to avoid the three loops in R?

Hi

I do not see any solution without loops but maybe others find it.

I think that you can do it in one loop. Best structure for loop will be list.

In each cycle you will compute matrix with diagonal NA

mat<-matrix(1,nrow=number of items, ncol=number of items)
diag(mat) <- NA

apply(price chunk * mat, 2, mean, na.rm=T)

So when I named your example as temp I get

fac <- interaction(factor(temp$Product), factor(temp$Year_Month), drop=T)
lll <- split(temp, fac)
for( i in 1:length(lll)) {
 mat <- matrix(1,  nrow=nrow(lll[[i]]), ncol= nrow(lll[[i]]))
 diag(mat) <- NA
 lll[[i]]$others <- apply(mat*lll[[i]][,3], 2, mean, na.rm=T)
 }

How to avoid the three loops in R?

Thread (5 messages)