zerodist: find point pairs with equal spatial coordinates

View source: R/zerodist.R

zerodistR Documentation

find point pairs with equal spatial coordinates


find point pairs with equal spatial coordinates


zerodist(obj, zero = 0.0, unique.ID = FALSE, memcmp = TRUE) 
zerodist2(obj1, obj2, zero = 0.0, memcmp = TRUE) 
remove.duplicates(obj, zero = 0.0, remove.second = TRUE, memcmp = TRUE)



object of, or extending, class SpatialPoints


object of, or extending, class SpatialPoints


object of, or extending, class SpatialPoints


distance values less than or equal to this threshold value are considered to have zero distance (default 0.0); units are those of the coordinates for projected data or unknown projection, or km if coordinates are defined to be longitude/latitude


logical; if TRUE, return an ID (integer) for each point that is different only when two points do not share the same location


use memcmp to find exactly equal coordinates; see NOTE


logical; if TRUE, the second of each pair of duplicate points is removed, if FALSE remove the first


zerodist and zerodist2 return a two-column matrix with in each row pairs of row numbers with identical coordinates; a matrix with zero rows is returned if no such pairs are found. For zerodist, row number pairs refer to row pairs in obj. For zerodist2, row number pairs refer to rows in obj and obj2, respectively. remove.duplicates removes duplicate observations if present, and else returns obj.


When using kriging, duplicate observations sharing identical spatial locations result in singular covariance matrices. This function may help identify and remove spatial duplices. The full matrix with all pair-wise distances is not stored; the double loop is done at the C level.

When unique.ID=TRUE is used, an integer index is returned. sp 1.0-14 returned the highest index, sp 1.0-15 and later return the lowest index.

When zero is 0.0 and memcmp is not FALSE, zerodist uses memcmp to evaluate exact equality of coordinates; there may be cases where this results in a different evaluation compared to doing the double arithmetic of computing distances.


# pick 10 rows
n <- 10
ran10 <- sample(nrow(meuse), size = n, replace = TRUE)
meusedup <- rbind(meuse, meuse[ran10, ])
coordinates(meusedup) <- c("x", "y")
zd <- zerodist(meusedup)
sum(abs(zd[1:n,1] - sort(ran10))) # 0!
# remove the duplicate rows:
meusedup2 <- meusedup[-zd[,2], ]
meusedup3 <- subset(meusedup, !(1:nrow(meusedup) %in% zd[,2]))
coordinates(meuse) <- c("x", "y")
zerodist2(meuse, meuse[c(10:33,1,10),])

sp documentation built on June 7, 2022, 1:10 a.m.