[Rcpp-devel] bug with DatetimeVector ?

[stat quant]

Right 2 lines were missing:

C++:

#include <Rcpp.h>
using namespace Rcpp;

//[[Rcpp::export]]
void bug(DataFrame df){
DatetimeVector dt = df["dt"];
}

R:

library("Rcpp")
sourceCpp("rcpp2.cpp")
df <-
data.frame(x=c(1L,NA,2L), dt=as.POSIXct(c(1L,NA,2L),origin='1970-01-01'))
print(df)
bug(df)
##here R crashes##

2013/7/19 Dirk Eddelbuettel <edd at debian.org>

>
> On 19 July 2013 at 14:05, stat quant wrote:
> | Hello list, I might have found a bug, would you mind checking it too?
>
> No, you didn't. See below.
>
> | In R:
> |
> | library("Rcpp")
> | sourceCpp("rcpp.cpp")
> | df = data.frame(x=c(1L,NA,2L),
> dt=as.POSIXct(c(1L,NA,2L),origin='1970-01-01'))
> |
> | df
> |    x                  dt
> | 1  1 1970-01-01 00:00:01
> | 2 NA                <NA>
> | 3  2 1970-01-01 00:00:02
> |
> | In C++
> |
> | //[[Rcpp::export]]
> | void bug(DataFrame df){
> |     DatetimeVector dt = df["dt"];
> | }
>
> That __obviously__ does not compile (no header, no namespace, ...) so I
> think
> you are being sloppy.
>
> Execution of a __repaired and reproducible__ example, first via littler
> then
> via R:
>
>
> edd at don:~$ r -lRcpp -e'sourceCpp("/tmp/statquant.cpp")'
>
> > df <- data.frame(x=c(1L,NA,2L),
> +                  dt=as.POSIXct(c(1L,NA,2L),origin='1970-01-01'))
>
> > print(df)
>    x                  dt
> 1  1 1969-12-31 18:00:01
> 2 NA                <NA>
> 3  2 1969-12-31 18:00:02
>
> > bug(df)
>
> > cat("Hello world\n")
> Hello world
> edd at don:~$ R -q -e 'Rcpp::sourceCpp("/tmp/statquant.cpp")'
> R> Rcpp::sourceCpp("/tmp/statquant.cpp")
>
> R> df <- data.frame(x = c(1, NA, 2), dt = as.POSIXct(c(1,
> +     NA, 2), origin = "1970-01-01"))
>
> R> print(df)
>    x                  dt
> 1  1 1969-12-31 18:00:01
> 2 NA                <NA>
> 3  2 1969-12-31 18:00:02
>
> R> bug(df)
>
> R> cat("Hello world\n")
> Hello world
> R>
> R>
> edd at don:~$
>
> A complete cpp snippet follows below.  Please construct better examples
> that
> actually __work__ next time, test them on __fresh sessions__ and consider
> including the R test code in the C++ snippet as I do here.  Oh, and stop
> using the hidous '=' for assignment in R ;-)
>
> Dirk
>
>
> -----------------------------------------------------------------------------
>
> #include <Rcpp.h>
>
> using namespace Rcpp;
>
> //[[Rcpp::export]]
> void bug(DataFrame df){
>     DatetimeVector dt = df["dt"];
> }
>
> /*** R
> df <- data.frame(x=c(1L,NA,2L),
>                  dt=as.POSIXct(c(1L,NA,2L),origin='1970-01-01'))
> print(df)
> bug(df)
> cat("Hello world\n")
> */
>
> -----------------------------------------------------------------------------
>
>
>
> |
> | Executing bug(df) in R crashes the R session on my box
> |
> | sessionInfo()
> | R version 3.0.0 (2013-04-03)
> | Platform: x86_64-w64-mingw32/x64 (64-bit)
> | locale:
> | [1] LC_COLLATE=French_France.1252  LC_CTYPE=French_France.1252
> |  LC_MONETARY=French_France.1252 LC_NUMERIC=C
> | LC_TIME=C
> | attached base packages:
> | [1] stats     graphics  grDevices datasets  utils     methods   base
> |
> | other attached packages:
> | [1] Rcpp_0.10.4  vimcom_0.9-8
> | loaded via a namespace (and not attached):
> | [1] tools_3.0.0
> |
> | Regards
> |
> | ----------------------------------------------------------------------
> | _______________________________________________
> | Rcpp-devel mailing list
> | Rcpp-devel at lists.r-forge.r-project.org
> | https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
> --
> Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/rcpp-devel/attachments/20130719/edd28e7e/attachment.html>