R help - Jul 2013 - Change R options only inside a package

Hi,

What should I do if I want to use "non-default" R-options for the
functions inside a package, but not affect the options of the rest of
the session? Specifically, I'm thinking of using "sum contrasts"
instead
of the default, in functions like "lm", etc. when they are called by
other functions in a package.

I can think on the following solutions:

1) Quick-and-dirty solution: use options() to modify the default
contrasts before calling my functions. If I don't want to affect the
rest of the session, I should reset the options before returning,
preferably with exception handling to avoid that the options are not
reset if my function fails.

2) Complex-but-tidier solution: Use function-specific arguments to tell
the contrasts of the factors that I'm using. E.g., I should create a
named list of contrasts and pass it as the argument "contrasts" every
time I call "lm":

mod <- lm(form, contrasts=list(factor1="contr.sum",
factor2="contr.sum"))


3) Encapsulated solution: create package-specific (unexported) versions
of "lm" and other functions, that already implement #2, e.g.:

lm <- function(...)
{
# Code that examines dots, to get the factors of the model, and create
the list of contrasts
# Then call the stats lm (be sure that dots did not already include a
"contrasts" element).
stats::lm(..., contrasts = list.of.factors.with.sum.contrasts)
}

I don't like #1, because changing the options back and forth may be
error-prone and lead to unexpected behaviour.

#2 is better, but prone to programmer errors: I should be careful of
not forgetting to set the "contrasts" argument every time I call
"lm".

I prefer #3, because it avoids the disadvantages of the previous
solutions, but perhaps it is overkill, hence my question: Is there a
"clean" way of setting the R-session options in such a way that they
only affect the functions and data inside a package?

Thanks and best regards

Helios De Rosario

INSTITUTO DE BIOMEC?NICA DE VALENCIA
Universidad Polit?cnica de Valencia ? Edificio 9C
Camino de Vera s/n ? 46022 VALENCIA (ESPA?A)
Tel. +34 96 387 91 60 ? Fax +34 96 387 91 69
www.ibv.org

  Antes de imprimir este e-mail piense bien si es necesario hacerlo.
En cumplimiento de la Ley Org?nica 15/1999 reguladora de la Protecci?n
de Datos de Car?cter Personal, le informamos de que el presente mensaje
contiene informaci?n confidencial, siendo para uso exclusivo del
destinatario arriba indicado. En caso de no ser usted el destinatario
del mismo le informamos que su recepci?n no le autoriza a su divulgaci?n
o reproducci?n por cualquier medio, debiendo destruirlo de inmediato,
rog?ndole lo notifique al remitente.

On 13-07-15 5:27 AM, Helios de Rosario wrote:
> Hi,
>
> What should I do if I want to use "non-default" R-options for the
> functions inside a package, but not affect the options of the rest of
> the session? Specifically, I'm thinking of using "sum
contrasts" instead
> of the default, in functions like "lm", etc. when they are called
by
> other functions in a package.
>
> I can think on the following solutions:
>
> 1) Quick-and-dirty solution: use options() to modify the default
> contrasts before calling my functions. If I don't want to affect the
> rest of the session, I should reset the options before returning,
> preferably with exception handling to avoid that the options are not
> reset if my function fails.
>
> 2) Complex-but-tidier solution: Use function-specific arguments to tell
> the contrasts of the factors that I'm using. E.g., I should create a
> named list of contrasts and pass it as the argument "contrasts"
every
> time I call "lm":
>
> mod <- lm(form, contrasts=list(factor1="contr.sum",
> factor2="contr.sum"))
>
>
> 3) Encapsulated solution: create package-specific (unexported) versions
> of "lm" and other functions, that already implement #2, e.g.:
>
> lm <- function(...)
> {
> # Code that examines dots, to get the factors of the model, and create
> the list of contrasts
> # Then call the stats lm (be sure that dots did not already include a
> "contrasts" element).
> stats::lm(..., contrasts = list.of.factors.with.sum.contrasts)
> }
>
> I don't like #1, because changing the options back and forth may be
> error-prone and lead to unexpected behaviour.
>
> #2 is better, but prone to programmer errors: I should be careful of
> not forgetting to set the "contrasts" argument every time I call
"lm".
>
> I prefer #3, because it avoids the disadvantages of the previous
> solutions, but perhaps it is overkill, hence my question: Is there a
> "clean" way of setting the R-session options in such a way that
they
> only affect the functions and data inside a package?
>
I'd recommend that in your own functions that need this, you put this at 
the beginning:

  saveOpts <- options(contrasts = c("contr.sum",
"contr.sum"))
  on.exit(options(saveOpts))

This means that the options will be changed just for the duration of the 
call.  If there are situations where users might not want this new 
default, make the new value a default of a function argument.

Duncan Murdoch


> Thanks and best regards
>
> Helios De Rosario
>
> INSTITUTO DE BIOMEC?NICA DE VALENCIA
> Universidad Polit?cnica de Valencia ? Edificio 9C
> Camino de Vera s/n ? 46022 VALENCIA (ESPA?A)
> Tel. +34 96 387 91 60 ? Fax +34 96 387 91 69
> www.ibv.org
>
>    Antes de imprimir este e-mail piense bien si es necesario hacerlo.
> En cumplimiento de la Ley Org?nica 15/1999 reguladora de la Protecci?n
> de Datos de Car?cter Personal, le informamos de que el presente mensaje
> contiene informaci?n confidencial, siendo para uso exclusivo del
> destinatario arriba indicado. En caso de no ser usted el destinatario
> del mismo le informamos que su recepci?n no le autoriza a su divulgaci?n
> o reproducci?n por cualquier medio, debiendo destruirlo de inmediato,
> rog?ndole lo notifique al remitente.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>