spatSample - stratified sampling function generating core dump error

[M3nin0]

Hi,

Thank you for the terra package!

I'd like to report an error I'm encountering when using terra::spatSample with method = 'stratified'. The issue arises when I run the following code:

# Example file available here: https://github.com/e-sensing/sitsdata/raw/refs/heads/main/inst/extdata/world-cover-mato-grosso/WORLD-COVER_2021_S12W057_class_2021-12-31.tif

rast <- terra::rast('https://github.com/e-sensing/sitsdata/raw/refs/heads/main/inst/extdata/world-cover-mato-grosso/WORLD-COVER_2021_S12W057_class_2021-12-31.tif')
points <- terra::spatSample(
    x = rast, 
    size = 500, 
    method = "stratified", 
    as.points = TRUE
)

I receive the following error message:

 *** caught segfault ***
address (nil), cause 'unknown'

Traceback:
 1: .External(list(name = "CppMethod__invoke_notvoid", address = <pointer: 0x55618601ad00>,     dll = list(name = "Rcpp", path = "/usr/local/lib/R/site-library/Rcpp/libs/Rcpp.so",         dynamicLookup = TRUE, handle = <pointer: 0x55618752b580>,         info = <pointer: 0x55618522fbe0>, forceSymbols = FALSE),     numParameters = -1L), <pointer: 0x55618940d770>, <pointer: 0x556187578660>,     .pointer, ...)
 2: x@pntr$sampleStratifiedCells(size, each, replace, .seed(), spatOptions())
 3: .sampleCellsStratified(x, size, each, replace, ext)
 4: sampleStratified2(x, size, replace = replace, as.df = as.df,     as.points = as.points, values = values, cells = cells, xy = xy,     ext = ext, warn = warn, each = each)
 5: .local(x, ...)
 6: terra::spatSample(rast, 500, "stratified", as.points = TRUE)
 7: terra::spatSample(rast, 500, "stratified", as.points = TRUE)
An irrecoverable exception occurred. R is aborting now ...
Segmentation fault (core dumped)

I was able to track this Segmentation fault (core dumped) when running the code in a terminal. When I run it in RStudio, the session crashes and R restarts. While testing, I tried other files - for example, this one - and I was able to reproduce the error.

The interesting point is that it doesn't throw the error every time (maybe related to the random sampling?). Sometimes it runs successfully and returns results, and other times it fails with the error:

> rast <- terra::rast('WORLD-COVER_2021_S12W057_class_2021-12-31.tif')
> points <- terra::spatSample(
    x = rast, 
    size = 500, 
    method = "stratified", 
    as.points = TRUE
)
> points
 class       : SpatVector 
 geometry    : points 
 dimensions  : 4000, 1  (geometries, attributes)
 extent      : -55.80254, -55.19904, -11.80196, -11.49587  (xmin, xmax, ymin, ymax)
 coord. ref. : lon/lat WGS 84 (EPSG:4326) 
 names       : cover
 type        : <num>
 values      :    10
                  10
                  10

Since the error doesn't appear consistently, I created a simple for loop to help reproduce it:

for (i in seq_len(100)) {
  print(paste0('Iteration: ', i))

  rast <- terra::rast('WORLD-COVER_2021_S12W057_class_2021-12-31.tif')
  points <- terra::spatSample(
    x = rast,
    size = 500,
    method = "stratified",
    as.points = TRUE
 )
}

I ran the same code on macOS Sequoia 15.5 and Ubuntu 22.04, using R versions 4.4.1 and 4.5.1, and was able to reproduce the issue.

We initially identified this behavior because we use terra as part of some sits functions for sampling data.

Please let me know if I can provide more information or other details.

Thanks a lot!

@M3nin0 / @OldLipe

session info - R 4.4.1

> sessionInfo()
R version 4.4.1 (2024-06-14)
Platform: x86_64-pc-linux-gnu
Running under: Ubuntu 22.04.3 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3 
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.20.so;  LAPACK version 3.10.0

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

time zone: Etc/UTC
tzcode source: system (glibc)

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] terra_1.8-54

loaded via a namespace (and not attached):
[1] compiler_4.4.1    tools_4.4.1       rstudioapi_0.17.1 Rcpp_1.0.14       codetools_0.2-20 

session info - R 4.5.1

> sessionInfo()
R version 4.5.1 (2025-06-13)
Platform: x86_64-pc-linux-gnu
Running under: Ubuntu 22.04.5 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.10.0 
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.10.0  LAPACK version 3.10.0

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

time zone: Etc/UTC
tzcode source: system (glibc)

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
[1] compiler_4.5.1   Rcpp_1.0.14      codetools_0.2-20 terra_1.8-54

[rhijmans]

I have not been able to reproduce this with the development version and the code below, using seeds 1 to 5

library(terra)
r <-  terra::rast('https://github.com/e-sensing/sitsdata/raw/refs/heads/main/inst/extdata/world-cover-mato-grosso/WORLD-COVER_2021_S12W057_class_2021-12-31.tif')

set.seed(5)
for (i in seq_len(100)) {
	print(i); flush.console();
	points <- terra::spatSample(r, size = 500, method = "stratified", as.points = TRUE)
}

[M3nin0]

Hi @rhijmans,

Thanks a lot for your support.

We have been attempting to reproduce the error using different files and scenarios. During these tests, we observed that the error appears to occur on different machines, depending on the size of the raster.

For example, on a Mac running macOS Sequoia 15.5 with 32 GB of RAM and R 4.4.1, the error occurs more frequently when using terra::spatSample on rasters with dimensions of 10000x10000 or larger. For this case, we created the following code:

set.seed(5)

nrows <- 10000
ncols <- 10000

example <- terra::rast(nrows = nrows, ncols = ncols)

terra::values(example) <- sample(1:10, size = nrows * ncols, replace = TRUE)

points <- terra::spatSample(
  x         = example,
  size      = 200,
  method    = "stratified",
  as.points = TRUE
)

We also ran this code in a Ubuntu 22.04, which has 22 GB of RAM and R 4.5.1 and the error also occurred.

However, when we ran the code above on a machine with Ubuntu 22.04 and 252 GB of RAM, we were not able to reproduce the error. We then modified the code to process a larger raster, and the error occurred:

set.seed(5)

nrows <- 50000
ncols <- 50000

example <- terra::rast(nrows = nrows, ncols = ncols)

terra::values(example) <- sample(1:10, size = nrows * ncols, replace = TRUE)

points <- terra::spatSample(
  x         = example,
  size      = 200,
  method    = "stratified",
  as.points = TRUE
)

For a real case, we also tested a raster with dimensions we typically have in our cases in sits, and the error occurred in all environments.

set.seed(777)

rast <- terra::rast("https://github.com/e-sensing/sitsdata/raw/refs/heads/main/inst/extdata/Rondonia-Class-2022-Mosaic/SENTINEL-2_MSI_MOSAIC_2022-01-05_2022-12-23_class_mosaic.tif")

points <- terra::spatSample(
    x         = rast,
    size      = 200,
    method    = "stratified",
    as.points = TRUE
)

We also tested the development version (9782806) and we were able to reproduce the error.

Thanks a lot!

@M3nin0 / @OldLipe