[R] error related to approxfun in R 2.7.0

Karl Marx tammisiech at gmail.com
Sun Jun 22 22:59:49 CEST 2008


Fantastic, works perfectly!
Would you have any pointers to where I can learn more about why/how this works?
Thanks!

Regarding the P.S., yes...  it seemed appropriate at the time :-)

2008/6/21, Martin Maechler <maechler at stat.math.ethz.ch>:
>>>>>> "KM" == Karl Marx <tammisiech at gmail.com>
>>>>>>     on Thu, 19 Jun 2008 23:01:31 +0200 writes:
>
>     KM> I'm testing R version 2.7.0 on windows and there seems
>     KM> to be a compatibility issue with objects that were
>     KM> created by "approxfun" in older versions. As long as the
>     KM> objects were created in version 2.7.0 things work ok,
>     KM> but calling the interpolated functions from R version
>     KM> 2.0.1 causes this error:
>
>     KM> Error in .C("R_approx", as.double(x), as.double(y),
>     KM> as.integer(n), xout = as.double(v), : C symbol name
>     KM> "R_approx" not in DLL for package "base"
>
>     KM> What's the problem and how can t be resolved (need to be
>     KM> compatible with old data)?
>
> approxfun() wa moved from "base" to "stats".
>
> If   f <- approxfun(...)  in R 2.0.1
> and you save f , load it in R 2.7.0,
> you can do
>
>     body(f)[[2]][["PACKAGE"]] <- "stats"
>
> and then use f(..) in R 2.7.0.
>
> Martin Maechler, ETH Zurich
>
> PS: Do I guess correctly, that you are Swiss?
>



More information about the R-help mailing list