You say "typical size of nodes --> 200000x2"
Do you mean that nodes has many many rows? Or many many columns? Or
both?
This is minimalist coding, but I'm not sure how fast it will run on your
data
aggregate(w, as.data.frame(nodes), sum)
Jean
"Weiser, Constantin" <constantin.wei...@uni-mainz.de> wrote on 06/28/2012
08:06:27 AM:
> Hi, all together. I have - a maybe trivial - problem with aggregating a
> list of weights.
>
> Here is the problem:
> - At first I have set of nodes (X/Y-coordinates) and associated
weights,
> where the set
> of nodes is typically not unique
> - I want to get a set of unique nodes and the sum of associated weights
>
> I am grateful for any help
>
>
> See for example:
>
> # weights:
> w <- c(1, 1, 1, 1, 1)
>
> # not unique set of nodes (X/Y-coordinates):
> nodes <- matrix(c(1,2,3,4,5,6,1,2,1,4),ncol=2, byrow=T)
>
>
> desired Result:
>
> #nodes
> [,1] [,2]
> [1,] 1 2
> [2,] 3 4
> [3,] 5 6
> [4,] 1 4
>
>
>
> #weights
> 2 1 1 1
>
>
>
> That is my solution, but it is very slow (typical size of nodes -->
> 200000x2):
>
> weights <- c(1, 1, 1, 1, 1)
> nodes <- matrix(c(1,2,3,4,5,6,1,2,1,4),ncol=2, byrow=T)
>
>
> ## to be replaced by a faster code
> drop.index <- duplicated(nodes)
> n.unique <- nodes[!drop.index, ]
> w.unique <- numeric(length(n.unique[,1]))
>
> lw <- length(weights)
> for (i in seq_along(w.unique)){
> index <-
as.logical(2==rowSums(nodes==matrix(rep(n.unique[i,],lw),byrow
> = TRUE, nrow=lw)))
> w.unique[i] <- sum(weights[index])
> }
> ##
>
> n.unique
> w.unique
>
>
>
>
>
> ^
> | X
> | /
> | /eiser, Constantin
> | / Gutenberg University of Mainz, Germany
> | * /\ / Chair of Statistics & Econometrics
> | \ / \ / Jakob-Welder-Weg 4, 55128 Mainz
> | \/ \/ House of Law and Economics II, Room 00-116
> | Tel: 0049 6131 39 22715
> +--------------------------------------------------------->
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
