cleaning up SpatialLines

-----Oorspronkelijk bericht-----
Van: Dylan Beaudette [mailto:dylan.beaudette at gmail.com]
Verzonden: donderdag 4 augustus 2011 0:46
Aan: ONKELINX, Thierry
CC: r-sig-geo at stat.math.ethz.ch
Onderwerp: Re: [R-sig-Geo] cleaning up SpatialLines

On Wed, Aug 3, 2011 at 3:11 PM, ONKELINX, Thierry
<Thierry.ONKELINX at inbo.be> wrote:

Dear all,

I have a SpatialLinesDataFrame containing a lot of short line fragments. Many

of them have exactly one other linefragment at one (or both) of their end
vertices. I would like to merge those linefragments to that each line is at both
end vertices either connected to at least two other lines or to no other line.

Does someone has an effecient solution for this? Via R or via grass?

I managed already to add a grouping variable to the data.frame. But I still

need to find a way to merge the coordinates of the lines from the same Group.
See the example below.

Best regards,

Thierry

library(sp)
Raw <-
SpatialLines(list(Lines(list(Line(cbind(x = c(0, 0, 1), y = c(0, 1,
1)))), ID = 1), Lines(list(Line(cbind(x = c(2, 2, 1), y = c(0, 1,
1)))), ID = 2), Lines(list(Line(cbind(x = c(-2, -2), y = c(0, -1)))),
ID = 3)
))
Raw <- SpatialLinesDataFrame(Raw, data.frame(ID = 1:3, Group =
c(1,1,2)))

Result <-
? ? ? ?SpatialLines(list(Lines(list(Line(cbind(x = c(0, 0, 1, 2, 2), y
= c(0, 1, 1, 1, 0)))), ID = 1),
? ? ? ? ? ? ? ? ? ? ? ?Lines(list(Line(cbind(x = c(-2, -2), y = c(0,
-1)))), ID = 2)
? ? ? ? ? ? ? ?))
Result <- SpatialLinesDataFrame(Result, data.frame(ID = 1:2, Group =
c(1,2)))

Would something like this work:

# using your example from above ...
library(rgeos)

# merge lines by group, manually
Raw.u <- gLineMerge(Raw[Raw$Group == 1, ])

# automated approach, using spatial intersection as the criteria for splitting
Raw.u <- gLineMerge(Raw)

... note that IDs are lost in the second, automated approach.
Iterating over IDs might be the safest approach.

Dylan