is.na returns TRUE for new advectors

[lawlerem]

If x is an advector manually created with advector, then is.na(x) will unexpectedly return TRUE regardless of the actual value of x. I encountered this when working with rasters/matrices with some NA entries where I needed to update all cells that were not NA. Hopefully this turns out to be an easy fix because this bug does not affect advectors coming from the parameter list.

As a (cumbersome) workaround you can create a non-advector copy of x and run is.na(notad_x).

Thanks for looking into this!

library(RTMB)

# is.na(x) = TRUE, unexpected
is.na(advector(1))
is.na(AD(1, force = TRUE))

# is.na(x) returns TRUE for all entries of x
# Overwritten x is
# [NA 0 0]
# [0  0 0]
# [0  0 0],
# unexpected
x<- advector(matrix(c(NA, rep(0, 8)), nrow = 3, ncol = 3))
x[!is.na(x)]<- seq_along(x[!is.na(x)])
x

# Workaround to get overwritten x to be
# [NA 3 6]
# [1  4 7]
# [2  5 8],
# as expected
notad_x<- matrix(c(NA, rep(0, 8)), nrow = 3, ncol = 3)
x<- advector(notad_x)
x[!is.na(notad_x)]<- seq_along(notad_x[!is.na(notad_x)])
x

# is.na(x) = FALSE, expected
obj<- MakeADFun(
    function(pars) {
        print(is.na(pars[[1]]))
        return(0)
    },
    list(1)
)

Session Info:

R version 4.4.1 (2024-06-14)
Platform: x86_64-pc-linux-gnu
Running under: Pop!_OS 22.04 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

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

time zone: America/Halifax
tzcode source: system (glibc)

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

other attached packages:
[1] RTMB_1.5

loaded via a namespace (and not attached):
[1] MASS_7.3-61      compiler_4.4.1   Matrix_1.6-5     TMB_1.9.14       Rcpp_1.0.13      codetools_0.2-19
[7] grid_4.4.1       lattice_0.22-5  

[kaskr]

Thanks @lawlerem Should be fixed now. Note that the same issue applies for is.nan is.finite etc.

[lawlerem]

Great, thank you!