Switching between two CRS in R (with rgdal)

Question

I'm (trying to) do operations on pairs of geographical points. I have the coordinates of my points in WGS84, and I need to have them in the Lambert 2 Extended CRS (LIIE). I'm trying to do it using rgdal.

Here's what I'm doing :

library("rgdal")
library("sp")

# Loading CRS
WGS84<-"+proj=longlat +ellps=WGS84 +towgs84=0,0,0,-0,-0,-0,0 +no_defs"
LIIE<-"+proj=lcc +lat_1=46.8 +lat_0=46.8 +lon_0=0 +k_0=0.99987742 +x_0=600000 +y_0=2200000 +a=6378249.2 +b=6356515 +towgs84=-168,-60,320,0,0,0,0 +pm=paris +units=m +no_defs"

# Loading the pairs of points
matrix<-read.table(file="file_with_coordinates.csv", header=TRUE, sep=";", stringsAsFactors = F)

The columns of matrix are as follow : origin_id, destination_id, or_lon, or_lat, de_lon, de_lat. Obviously, only the last 4 columns need to be transformed from WGS84 to LIIE.

I'm able to transform the coordinates by doing this :

matrix_sp<-SpatialPoints(coords = od_matrix[,c("de_lon","de_lat","or_lon","or_lat")],proj4string = CRS(WGS84))
matrix_sp_liie<-spTransform(od_matrix_sp, CRSobj = CRS(LIIE))
matrix_liie<-data.frame(matrix_sp_liie)

However, I therefore lose the origin and destination IDs... (And I don't have anything left that could allow me to merge back together matrix_liie with the origin/destination ids in matrix_sp).

I tried this (it's basically the same code but with destination_id and oririgin_id included in the first line), but I couldn't really get to something interesting (I get a Error in .local(obj, ...) : cannot derive coordinates from non-numeric matrix error).

od_matrix_sp<-SpatialPoints(coords = od_matrix[,c("destination_id","oririgin_id","de_lon","de_lat","or_lon","or_lat")],proj4string = CRS(WGS84))
matrix_sp_liie<-spTransform(od_matrix_sp, CRSobj = CRS(LIIE))
matrix_liie<-data.frame(matrix_sp_liie)

Any idea on how I could achieve this ?

Thanks.

Sample from CSV :

origin_id    destination_id    or_lon    or_lat    de_lon     de_lat
123_a        005               3.88      45.6      1.56       46.7
123_b        006               5.10      41.1      2.4        42.6

Can you give a sample few lines from your csv file? Because I suspect the answer is just to create a `SpatialLinesDataFrame`, or a `SpatialPointsDataFrame` with twice the number of features as rows, and use the ID values to link them... — Spacedman, Jan 22 '16 at 14:55
Ah yes, you're right I didn't think to do that, I edited my answer. If you want to post this as an answer I'll withdraw mine — Victorp, Jan 22 '16 at 15:23
Thank you both for your comments. I added a sample from the CSV. I'm not sure what you mean with `twice the number of features as rows, and use the ID values to link them` ? — François M., Jan 25 '16 at 09:13

Victorp · Answer 1 · 2016-01-22T15:19:40.230

Hi it's sp that does the conversion, and you can do that without use SpatialPoints, just specify which columns in matrix are the coordinates with coordinates, here an example :

library("sp")
# Some coordinates
latlong <- data.frame(
  ID = 1:8,
  LETTERS = LETTERS[1:8],
  lon = runif(n = 8, min = 2.0798, max = 2.9931),
  lat = runif(n = 8, min = 48.6823, max = 49.0698)
)

# Loading CRS
WGS84<-"+proj=longlat +ellps=WGS84 +towgs84=0,0,0,-0,-0,-0,0 +no_defs"
LIIE<-"+proj=lcc +lat_1=46.8 +lat_0=46.8 +lon_0=0 +k_0=0.99987742 +x_0=600000 +y_0=2200000 +a=6378249.2 +b=6356515 +towgs84=-168,-60,320,0,0,0,0 +pm=paris +units=m +no_defs"

# Set which var are the coord
coordinates(latlong) <- c("lon", "lat")
# Set the proj
proj4string(latlong) <- CRS(WGS84)
# Convert
latlon_conv <- spTransform(x = latlong, CRSobj = CRS(LIIE))
# Back to data.frame
latlon_conv <- as.data.frame(latlon_conv)
latlon_conv # maybe change columns names...
#   ID LETTERS      lon     lat
# 1  1       A 632441.1 2440172
# 2  2       B 633736.7 2434332
# 3  3       C 586298.5 2411320
# 4  4       D 645107.6 2410351
# 5  5       E 642454.6 2443052
# 6  6       F 628371.7 2448833
# 7  7       G 625445.7 2436324
# 8  8       H 624509.7 2443864

EDIT : After seeing @Spacedman comment, you can effectively use SpatialPointsDataFrame instead of SpatialPoints :

latlong.sp <- SpatialPointsDataFrame(
  coords = latlong[, c("lon", "lat")],
  data = latlong[, c("ID", "LETTERS")],
  proj4string = CRS(WGS84)
)

latlon_conv <- spTransform(x = latlong.sp, CRSobj = CRS(LIIE))
latlon_conv.df <- as.data.frame(latlon_conv)
latlon_conv.df

I tried the `SpatialPointsDataFrame` solution, and it partially works : it converts the coordinates of one of the two points, but the other one stays in WGS84. — François M., Jan 25 '16 at 09:15

Switching between two CRS in R (with rgdal)

1 Answers1