[Rd] Including data in packages
Janko Thyson
janko.thyson at ku-eichstaett.de
Thu Oct 14 21:27:39 CEST 2010
Hi Dirk,
thanks for the quick reply. So all datasets, regardless if they should be
hidden from the user, should go into /data? If that's the convention, that's
fine with me. But the problem is that there is no data-directory in
R_HOME/library/2-11.1/mypackage after running R CMD build and R CMD INSTALL.
I've experimented with the options --binary and --use-zip-data but without
luck.
I think I don't understand the "optional directory" procedure in general
yet: don't I just need to dump stuff in /data, /inst/doc, /exec etc. and
have them "transferred" (sorry, I don't know a better word) to the tar.gz
and eventually to the installed package?
Regards,
Janko
> -----Ursprüngliche Nachricht-----
> Von: Dirk Eddelbuettel [mailto:edd at debian.org]
> Gesendet: Donnerstag, 14. Oktober 2010 21:18
> An: Janko Thyson
> Cc: r-devel at r-project. org
> Betreff: Re: [Rd] Including data in packages
>
>
> On 14 October 2010 at 21:06, Janko Thyson wrote:
> | The thing is that I don't understand how the actual 'transfer'
> happens when
> | the package is build and installed. Where do the objects 'live' after
> | installation and how can I access them after firing
> library(mypackage)? For
> | example, there is a 'data' directory for package 'tm'
> | (R_HOME/R-2.11.1/library/tm) but such a directory is missing even if
> I chose
> | option 3) (as stated above) to get my data included. Also, how can I
> | actually execute the lazyload?
>
> If in doubt, copy existing packages, preferably recommended ones:
>
> edd at max:~/src/debian/CRAN$ find -type d -name data | sort
> ./boot-1.2.42/data
> ./boot-1.2.43/data
> ./car-2.0-2/data
> ./cluster-1.13.1/data
> ./effects-2.0.10/data
> ./fEcofin-290.76/data
> ./fExtremes-2100.77/data
> ./fGarch-2110.80/data
> ./fImport-2110.79/data
> ./fPortfolio-2110.79/data
> ./gdata-2.8.0/data
> ./gplots-2.8.0/data
> ./gtools-2.6.2/data
> ./lattice-0.19-11/data
> ./lattice-0.19-13/data
> ./latticeExtra-0.6-14/data
> ./lme4-0.999375-35/data
> ./lmtest-0.9.27/data
> ./MASS-7.3-7/data
> ./Matrix-0.999375-44/data
> ./misc3d-0.7-1/data
> ./multcomp-1.2-3/data
> ./nlme-3.1.96/data
> ./nlme-3.1.97/data
> ./nws-2.0.0.3/data
> ./rggobi-2.1.16/data
> ./robustbase-0.5-0-1/data
> ./rpart-3.1.46/data
> ./sandwich-2.2-6/data
> ./sm-2.2-4.1/data
> ./sn-0.4-16/data
> ./strucchange-1.4-1/data
> ./strucchange-1.4-2/data
> ./survival-2.35-8/data
> ./timeSeries-2120.89/data
> ./tseries-0.10-22/data
> ./urca-1.2-4/data
> edd at max:~/src/debian/CRAN$
>
> The vote is clearly in favour of data/ over inst/data.
>
> Hth, Dirk
>
> --
> Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com
More information about the R-devel
mailing list