zerodist {gstat} | R Documentation |
find point pairs with equal spatial coordinates
zerodist(x, y, z, zero = 0.0)
x |
vector with x-coordinate |
y |
vector with y-coordinate (may be missing) |
z |
vector with z-coordinate (may be missing) |
zero |
value to be compared to for establishing when a distance is considered zero (default 0.0) |
pairs of row numbers with identical coordinates, numeric(0) if no such pairs are found
Duplicate observations sharing identical spatial locations result in singular covariance matrices in kriging situations. This function may help identifying spatial duplications, so they can be removed. A matrix with all pair-wise distances is calculated, so if x, y and z are large this function is slow
data(meuse) # pick 10 rows n <- 200 ran10 <- sample(nrow(meuse), size = n, replace = TRUE) meusedup <- rbind(meuse, meuse[ran10, ]) zd <- zerodist(meusedup$x, meusedup$y) sum(abs(zd[1:n,1] - sort(ran10))) # 0! # remove the duplicate rows: meusedup2 <- meusedup[-zd[,2], ]