argument to which is not logical

[mcol]

I've just come across this odd behaviour while attempting to remove columns containing only NAs from a data.table:

library(data.table)
dt <- data.table(iris)[, Petal.Length := NA]

dt[, which(sapply(.SD, function(z) all(is.na(z))))]            # works as expected
# Petal.Length
#            3

dt[, which(sapply(.SD, function(z) all(is.na(z)))) := NULL]    # error
# Error in which(sapply(.SD, function(z) all(is.na(z)))) :
#   argument to 'which' is not logical

There is probably a better way of achieving that, but in any case the error message is extremely puzzling!

# Output of sessionInfo()
R version 3.6.3 (2020-02-29)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.5 LTS

Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.7.1
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.7.1

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

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

other attached packages:
[1] data.table_1.14.0

loaded via a namespace (and not attached):
[1] compiler_3.6.3

[MichaelChirico]

.SD is not available on the LHS of :=

use dt instead; see #795

[mcol]

Would it be possible to mention that in the error message? As is it doesn't give any clue that that is the problem.

[MichaelChirico]

I'm afraid not. what's happening is that .SD is NULL until we populate it like normal. I suspect that it would be roughly equivalent effort to make .SD work on the LHS as it would be to give a helpful error in that case.

FWIW there is an open PR for fixing the referenced issue, so hopefully it can be fixed soon enough.