complete.cases does not work with nanotime

[vermosen]

Hi,

I have a problem with the following example:

library(nanotime)
library(data.table)

DT <- data.table(time=as.nanotime(seq(1, 3)))
DT[, shift.val := shift(time, 1, fill=NA)]

is.na(DT[1, shift.val])       # TRUE
DT[complete.cases(DT)]        # still shows up here

here is my current configuration:

> sessionInfo()
R version 4.0.0 (2020-04-24)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS Linux 8 (Core)

Matrix products: default
BLAS:   /opt/r40-gcc84/lib64/R/lib/libRblas.so
LAPACK: /opt/r40-gcc84/lib64/R/lib/libRlapack.so

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_1.13.0 nanotime_0.3.2

loaded via a namespace (and not attached):
[1] zoo_1.8-8       bit_4.0.4       compiler_4.0.0  RcppCCTZ_0.2.9
[5] Rcpp_1.0.5      bit64_4.0.5     grid_4.0.0      lattice_0.20-41

[vermosen]

Actually, this doesn't seems to relate to data.table. The same problem do occur with regular data.frame objects. There is a seemingly related remark in ?complete.cases ...

[MichaelChirico]

What is as.nanotime? I think nanotime(.) is intended

[MichaelChirico]

Thanks for the report, I can reproduce and I can confirm the note you saw in complete.cases is indeed the source of the issue:

A current limitation of this function is that it uses low level functions to determine lengths and missingness, ignoring the class.

The issue is that nanotime is stored (from R's perspective) as double, so complete.cases is looking for NA_real_ but the missing value for nanotime is different.

Luckily for you, there is a data.table method for na.omit that is bit64-aware, so this should work:

na.omit(DT)

Unluckily for you, it doesn't (yet) 😄

However, a fix is ready in #4752

[jangorecki]

I think we can close this one. It is base R issue which s already documented in base R.