[R-sig-Geo] help: Problem getting centroids inside lists
MacQueen, Don
m@cqueen1 @ending from llnl@gov
Wed Sep 12 22:43:05 CEST 2018
On any R mailing list, whenever you say "it doesn't work", you should always copy exactly what R command you gave, and any error messages.
First notice:
> ct$a$two
SpatialPoints:
lon lat
[1,] -18 -2
[2,] -16 50
Coordinate Reference System (CRS) arguments: NA
Then compare these:
> coordinates(ct$a$two)
lon lat
[1,] -18 -2
[2,] -16 50
> coordinates(slot(ct$a$two,'coords'))
lon lat
[1,] -18 -2
[2,] -16 50
The results are the same. Why are you using the slot() function? There is no need, and it makes it more difficult to understand.
You have assumed the mean() function will give you what you want when there are two points. It doesn't. Try it and see:
> mean( coordinates(ct$a$two))
[1] 3.5
> mean(coordinates(slot(ct$a$two,'coords')))
[1] 3.5
Replace the mean() function with a function that will give you the "centroid" of two points, however you want to define that centroid. Perhaps the means of the two columns? R has a function for that.
-----------------
In this bit:
ctply <- lapply(X = s, FUN = function(x) sapply(X = ct[[x]],
FUN = function(y)
Using X = s is wrong because you haven't defined or created s anywhere. X should be ct.
Also, X = ct[[x]] is wrong. When the first FUN is executed, it will be supplied automatically with the whole of ct$a, then ct$b. Try
ctply <- lapply(X = s, FUN = function(x) sapply(X = x,
FUN = function(y)
The whole thing is easier to understand and test if you define the functions outside the lapply and sapply calls.
-Don
--
Don MacQueen
Lawrence Livermore National Laboratory
7000 East Ave., L-627
Livermore, CA 94550
925-423-1062
Lab cell 925-724-7509
On 9/12/18, 11:42 AM, "R-sig-Geo on behalf of Ariel Fuentesdi" <r-sig-geo-bounces using r-project.org on behalf of ariel.fuentesdi using usach.cl> wrote:
I went to a more general approach, that is:
ctply <- lapply(X = s, FUN = function(x) sapply(X = ct[[x]],
FUN = function(y) if(length(y)>1) geosphere::centroid(slot(y,
"coords"))
else sp::coordinates(slot(y, "coords"))))
But now I want to add the case when they are only two elements. The dataset
will be:
ct <- list(a = list(one = data.frame(lon = c(-180, -160, -60), lat = c(-20,
5, 0)),
two = data.frame(lon = c(-18, -16), lat = c(-2, 50))),
b = list(one = data.frame(lon = c(-9, -8, -3), lat = c(-1, 25,
5)),
two = data.frame(lon = c(-90), lat = c(-1))))
coordinates(ct$a$one) <- ~lon+lat
coordinates(ct$a$two) <- ~lon+lat
coordinates(ct$b$one) <- ~lon+lat
coordinates(ct$b$two) <- ~lon+lat
And I did the following but it doesn't work:
ctply <- lapply(X = s, FUN = function(x) sapply(X = ct[[x]],
FUN = function(y)
if(length(y)>2) geosphere::centroid(slot(y, "coords"))
else if (length(y) == 1)
sp::coordinates(slot(y, "coords"))
else
mean(sp::coordinates(slot(y, "coords")))))
I need the result of each element of the list will be a matrix of two rows
per column (named: one, two). How do I fix it?
Regards,
Ariel
2018-09-10 17:29 GMT-03:00 MacQueen, Don <macqueen1 using llnl.gov>:
> If all of your data frames had enough points then this should work:
>
> myfun1 <- function(le) {
> list(one=geosphere::centroid(coordinates(le$one)),
> two=geosphere::centroid(coordinates(le$two))
> )
> }
>
> lapply(ct, myfun1)
>
> To handle the one point case, try the following, which I think works with
> your example data, but is untested if there just two points.
> It also assumes the data frames are named 'one' and 'two', and will ignore
> any others. To handle other names, I think you
> could replace $one with [[1]] and $two with [[2]] .
>
> myfun <- function(le) {
> list(one=if (length(le$one)>1) geosphere::centroid(coordinates(le$one))
> else coodinates(le$one),
> two=if (length(le$two)>1) geosphere::centroid(coordinates(le$two))
> else coordinates(le$two)
> )
> }
>
> lapply(ct, myfun)
>
>
> To see a little bit of what's going on, try
>
> myfun(ct[[1]])
> myfun1(ct[[1]])
> myfun1(ct[[2]])
> myfun(ct[[2]])
>
>
>
> I would also verify that the length method for objects of class
> SpatialPoints does return the number of points, as it appears to:
>
> > class(ct$a$two)
> [1] "SpatialPoints"
> attr(,"package")
> [1] "sp"
>
> > length(ct$a$one)
> [1] 3
> > length(ct$a$two)
> [1] 3
> > length(ct$a$two)
> [1] 3
> > length(ct$b$two)
> [1] 1
>
> --
> Don MacQueen
> Lawrence Livermore National Laboratory
> 7000 East Ave., L-627
> Livermore, CA 94550
> 925-423-1062
> Lab cell 925-724-7509
>
>
>
> On 9/10/18, 10:56 AM, "R-sig-Geo on behalf of Ariel Fuentesdi" <
> r-sig-geo-bounces using r-project.org on behalf of ariel.fuentesdi using usach.cl>
> wrote:
>
> Hi everyone,
>
> I have a list of coordinates called "ct" and I want to extract the
> centroids of each sublist, but it only works when it has only 3 or more
> points. In ct$b$two only has one point; in that case, I would rescue
> that
> point as If it were a centroid.
>
> This is what I did:
>
> library(dplyr)
> library(geosphere)
>
> ct <- list(a = list(one = data.frame(lon = c(-180, -160, -60), lat =
> c(-20,
> 5, 0)),
> two = data.frame(lon = c(-18, -16, -6), lat =
> c(-2, 50,
> 10))),
> b = list(one = data.frame(lon = c(-9, -8, -3), lat = c(-1,
> 25,
> 5)),
> two = data.frame(lon = c(-90), lat = c(-1))))
>
> coordinates(ct$a$one) <- ~lon+lat
> coordinates(ct$a$two) <- ~lon+lat
> coordinates(ct$b$one) <- ~lon+lat
> coordinates(ct$b$two) <- ~lon+lat
>
> s <- 1:length(ct)
> ctply <- list()
> for (i in s){
> for (j in 1:length(ct[[i]])) {
> ctply[[i]][j] <- ifelse(test = lengths(ct[[i]][1]) > 2, yes =
> sapply(X =
> ct[[i]][j],
> FUN = function(y) geosphere::centroid(slot(y, "coords"))), no =
> ct[[i]][j] )
> }
> }
>
> Thanks in advance
>
> Regards,
> Ariel
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>
>
>
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
R-sig-Geo using r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
More information about the R-sig-Geo
mailing list