I was about to produce some elevation transects when I noticed some weird behaviour I cannot explain, because the transects per se should not differ structurally. Using example data, t1 - the E-W-transect - produces 95 values using extract but t2 - the N-S-transect - only 2 values consisting of NA.
library(terra)
#> terra 1.6.7
filename <- system.file("ex/elev.tif", package="terra")
r <- rast(filename)
r
#> class : SpatRaster
#> dimensions : 90, 95, 1 (nrow, ncol, nlyr)
#> resolution : 0.008333333, 0.008333333 (x, y)
#> extent : 5.741667, 6.533333, 49.44167, 50.19167 (xmin, xmax, ymin, ymax)
#> coord. ref. : lon/lat WGS 84 (EPSG:4326)
#> source : elev.tif
#> name : elevation
#> min value : 141
#> max value : 547
# create transects
e <- ext(r)
t1 <- rbind(c(e[1], 49.8), c(e[2], 49.8)) |> vect("lines", crs = "epsg:4326")
t2 <- rbind(c(6.0, e[3]), c(6.0, e[4])) |> vect("lines", crs = "epsg:4326")
# inspect
plot(r)
plot(t1, add = TRUE)
plot(t2, add = TRUE)
# extract values by lines
t1_vals <- terra::extract(r, t1)
dim(t1_vals)
#> [1] 95 2
head(t1_vals)
#> ID elevation
#> 1 1 NA
#> 2 1 NA
#> 3 1 NA
#> 4 1 NA
#> 5 1 402
#> 6 1 418
t2_vals <- terra::extract(r, t2)
dim(t2_vals)
#> [1] 2 2
head(t2_vals)
#> ID elevation
#> 1 1 NA
#> 2 1 NA
I would expect a vector of length 90 for t2_vals$elevation, according to nrow(r).
What am I missing?
Edit:
touches = TRUE does not change the result returned by extract(), but exact = TRUE apparently does although I would not expect a change in behaviour here since y is not of polygon type (c.f. ?extract):
t1_vals <- terra::extract(r, t1, exact = TRUE)
dim(t1_vals)
#> [1] 95 3
head(t1_vals)
#> ID elevation fraction
#> 1 1 NA 0.01052632
#> 2 1 NA 0.01052632
#> 3 1 NA 0.01052632
#> 4 1 NA 0.01052632
#> 5 1 402 0.01052632
#> 6 1 418 0.01052632
t2_vals <- terra::extract(r, t2, exact = TRUE)
dim(t2_vals)
#> [1] 90 3
head(t2_vals)
#> ID elevation fraction
#> 1 1 NA 0.01111183
#> 2 1 NA 0.01111181
#> 3 1 515 0.01111179
#> 4 1 491 0.01111178
#> 5 1 469 0.01111176
#> 6 1 476 0.01111175